4. FISHERY CASE STUDIES

Transcription

1 Mediterranean 4. FISHERY CASE STUDIES The Mediterranean swordfish fishery Overview The Mediterranean Sea provides the earliest record of swordfish fishing, from about 1000 BC. Fishers from a diverse group of nations now use mainly longlines to catch swordfish in the Mediterranean. Driftnets are still used there, accounting for as much as 40 per cent of the Mediterranean swordfish catch. Total swordfish catches increased slowly from 1962 to about 5000 t per year by Catches then increased rapidly peaking at almost t in They then declined substantially, fluctuating between and t per year for the remainder of the 1990s. The catch levels are comparable to those of the North Atlantic, despite the Mediterranean Sea being a much smaller water body. The Mediterranean is quite different from the world s oceans, however, so care should be taken in applying conclusions on Mediterranean swordfish to other fisheries. The high productivity of the Mediterranean resource may be related to intense upwelling of deep, nutrient-rich waters in several areas. The Mediterranean might also provide extensive areas and prolonged seasons suitable for swordfish spawning. In addition to the decline in total catch levels, several fishery indicators suggest that Mediterranean swordfish are overexploited. The size composition of catches has declined so that the fishery now relies on only two or three, mostly immature, age classes. The size at maturity of Mediterranean swordfish is smaller than that observed for the Atlantic swordfish. Their genetic diversity is low when compared to swordfish in the southern Atlantic where fishing pressure has been comparatively low. The fishery s reliance on small swordfish and the fluctuations in recruitment reflect significant reductions in the size of the parent stock which might cause wide fluctuations in recruitment. Yet, it is difficult for the diverse nations fishing in the Mediterranean to agree on and enforce appropriate management measures for swordfish. The fishery Development The Mediterranean has a long history of swordfish fishing. The Romans and Greeks caught swordfish in the Mediterranean Sea before biblical times. The harpoon fishery in the Straits of Messina was first reported in the tenth century BC and a driftnet fishery has existed in the same area since at least 177 BC (Di Natale 1991). The largest producers of swordfish in the Mediterranean are now Italy (43% of the 1997 total catch), Morocco (33%) and Spain (7%; ICCAT in press; Table 3, Figure 7). Algeria, Cyprus, Greece, Malta, Tunisia and Turkey also have directed fisheries targeting swordfish. Italy Italy has always been one of the largest harvesters of swordfish in the Mediterranean. Since Roman times, swordfish fisheries have operated in the Straits of Messina, using harpoons, in the Straits of Sicily (driftnet and longline) and in the Tyrrhenian Sea (driftnet; Map 2). Italian fishers landed per cent of the total Mediterranean swordfish catch until 1990 (GFCM ICCAT 1996, p. 4). Its share of the catch then declined to about per cent of the total. Italy s fleet now fishes in most areas of the Mediterranean, including the Ligurian, Ionian and Adriatic Seas. Swordfish fishing commenced in those seas during the 1960s, initially using longlines. In the mid-1980s large-scale pelagic driftnets became popular, with Italy s government promoting them as having more desirable conservation qualities than other fishing gear (Folsom 1997b, p. 67). The modern driftnet fleet was extremely successful at catching swordfish and landings trebled by In 1992 international concern over the bycatch of marine mammals led the United Nations (UN) General Assembly to introduce a moratorium on the use of large-scale (>2.5 km long) pelagic driftnets in international Swordfish fisheries 35

2 Table 3. Summary of Mediterranean swordfish fleets. Fleet Total catch (1997) Fishing gear(s) Number of boats Boat size Fishing season Italy 6104 t longline 701 (1992) all year driftnet (1992) m, GRT harpoon 14 (1992) Morocco 4900 t longline 140 in total (1997) April August ~10 GRT September October (peak) driftnet Spain 1264 t longline 75 (1993) <20 GRT June November driftnet Algeria 807 t longline four-panel trawl Greece 750 t longline 400 (1994) <5 GRT May September Tunisia 346 t longline 40 (1997) 9 24m Turkey 320 t longline 1 5 GRT driftnet 1 5 GRT harpoon 1 5 GRT Malta 83 t longline 10 all year, up to 60 seasonally June August (peak) Cyprus 75 t longline 19 (1993) m March October waters. Since the peak catches of about t in 1988 and 1989, Italy s catches have declined, fluctuating between 4789 and 7765 t per year, mainly as a result of reduced driftnet effort. Morocco Morocco began fishing for swordfish in 1961 using mostly pelagic longlines (Folsom 1997b, pp ). Landings steadily increased, reaching 327 t in Then they declined, eventually ceasing in The fishery resumed in Catches were small (~50 t per year) until 1989, when driftnetting commenced and catches increased substantially. By 1997 Morocco was the second-largest harvester of swordfish, landing 33 per cent of the total swordfish catch in the Mediterranean (GFCM ICCAT 1998, p. 15). Spain Spanish fishers have longlined for swordfish in the Mediterranean for more than a century, but catch reporting did not commence until 1960 (Folsom 1997b, p. 102). Catches grew rapidly to average 1000 t per year during Part of the increase is probably due to increased reporting of catches. During catches dropped sharply to less than 100 t per year. They then slowly recovered to reach 1120 t in Since then, Spain s catches have remained stable, averaging about 1200 t per year. Algeria Algeria began fishing for swordfish in 1971 with pelagic longlines (Folsom 1997a, p. 86). By 1975 Algeria s annual catches of swordfish exceeded 500 t. The catch continued to increase each year, exceeding 2600 t in In 1988 Algerian fishers began targeting swordfish using locally developed four-panel trawls (chalut à quartre faces). Longline catches of swordfish eventually declined to about 30 per cent of the Mediterranean catch. 36 Bureau of Rural Sciences

3 Mediterranean Greece Greek fishers began using modern pelagic longlines (of a longer length than the traditional gear) to catch swordfish in However, landings data were not collected until Catches increased rapidly and the fleet landed about t of swordfish per year. Tunisia Tunisia began longlining for swordfish off its northern coastline in In the past, swordfish were taken mainly as a bycatch of Tunisia s driftnet and purse seine fisheries for tuna. In 1997 Tunisia s fishers reported a total catch of 346 t of swordfish. Turkey Turkish fishers have landed swordfish for many years. There are, however, no catch records prior to Turkey s catch of swordfish has fluctuated widely since the 1960s (Folsom 1997b, p. 164). From a peak of 326 t in 1966, catches declined to less than 10 t in 1976, increased to more than 590 t in 1988 and then decreased again to 100 t in Turkish fishers reported 320 t of swordfish in Malta Maltese fishers have been catching swordfish for many years, but the first catch (~100 t) was not reported until 1970 (Folsom 1997b, p. 156). Malta s swordfish catches then fluctuated between 71 and 222 t. Catches were generally lower after 1992, ranging from 42 t (1994) to 83 t (1997). Cyprus Spain Algeciras Strait of Gibraltar France Ligurian Sea Sardinia Balearic Islands Corsica Straits Tunisia Italy Tyrrhenian Sea of Sicily Croatia Sicily Malta Adriatic Sea Albania Greece Crete Aegean Sea BLACK SEA Turkey Cyprus Morocco Algeria Levant Sea 0 10 Libya 20 Egypt Map 2. The Mediterranean Sea, showing approximate location of 6 and 12 nm limits and the 2000 m isobath. Swordfish fishing began in Cyprus in 1974 following a visit by Greek and Italian longline fishers, who instructed local fishers on how to longline for swordfish (Folsom 1997b, p. 139). Longlining commenced immediately and the fishery grew rapidly. Algeria landed 59 t in 1976 and up to 100 t by There was a slump in catches in 1983 and 1984, the reasons for which are not clear. Since 1985 catches have fluctuated between 71 and 159 t per year. Other nations Libya reported t of swordfish per year during , after which the fishery apparently ceased (Folsom 1997a, p. 98). In recent years Libya has reported small amounts of swordfish as a bycatch of fishing for northern bluefin tuna. Swordfish fisheries 37

4 Several other nations also take swordfish as a bycatch of their northern bluefin tuna operations. The Japanese started longlining in the Mediterranean Sea in Japan s fleet targets northern bluefin tuna, with a bycatch of swordfish. 1 In the Mediterranean less than three per cent of Taiwan s longline catch is swordfish, amounting to one t in 1996 and three t in 1998 (GFCM ICCAT 1998, p. 13). Croatia lands swordfish caught as a bycatch of its northern bluefin tuna operations and also by subsistence fishers and recreational anglers. However, there is no catch monitoring; based on the number of boats, Croatia s swordfish landings are estimated to be less than ten t per year. Small French boats that target northern bluefin tuna offshore land swordfish as a bycatch in small ports where no monitoring occurs so the amount is not known. In recent years there have been many sightings of ghost ship longliners that operate without flags (or any identification) in the Mediterranean Sea. They are believed to target northern bluefin tuna, often during the closed season. Little is known about their catch levels or bycatch (GFCM ICCAT 1998, p. 9). Catches Landings 3 ('000 t) Italy Morroco Sp ain Algeria Greece Tunisia Turkey Malta Cyprus Albania Others Figure 7. Annual landings of swordfish by the ten leading harvesters in the Mediterranean Sea, 1997 (Dr P. Kebe, 12 May 1999). Catch trends Since the late 1980s the Mediterranean swordfish resource has supported catch levels comparable to those of the North Atlantic, despite the Mediterranean Sea being a much smaller water body. The high productivity of the Mediterranean resource may be related to the extensive areas suitable to swordfish for spawning. Alternatively, it might be due to the small size and age at first maturity of Mediterranean swordfish. Furthermore, throughout the year oceanographic conditions in many areas are suitable for swordfish spawning (ICCAT in press a). Mediterranean swordfish catches showed a gradual upward trend from 1962 to 1972, then stabilised at about 5000 t until 1983 (Figure 8). Catches increased sharply after 1983, peaking at t in The catch increases were partly due to increased pelagic driftnet fishing and improved data collection and reporting by several nations. 1 The catch of swordfish is around one per cent of the total catch by weight (Takeuchi 1996, p. 108). 38 Bureau of Rural Sciences

5 Mediterranean Since 1988 Mediterranean swordfish catches have declined and for 1990 and 1991 fell below t. The decline represents a reduction of nearly 45 per cent in the total catch of swordfish over just two years. Catches fluctuated between and t per year for the remainder of the 1990s. Recently, there has been an increase in the proportion of surface (harpoon and driftnet) catches of swordfish (GFCM ICCAT 1998). Size composition Since the late 1980s small juvenile swordfish (<120 cm) have dominated the catches of most Mediterranean swordfish fisheries. During more than half of Greece s swordfish catch were juveniles that had never spawned (Megalofonou et al. 1990, p. 171). By 1994, 64 per cent of the total Mediterranean swordfish catch (by number) and more than 70 per cent of the longline catch, were juvenile swordfish that had never spawned (GFCM ICCAT 1996, p. 64). The proportion of small swordfish taken by longlines is much higher than that taken with driftnet or other fishing gear. Species associations A 1997 survey of Spain s landings revealed that shark comprise more than 90 per cent of the total longline catch in the western Mediterranean Sea, with the blue shark accounting for 82 per cent (Buencuerpo et al. 1998). Swordfish comprise only 8 per cent of the total catch. In contrast, Longline Other Catch ('000 t) Year 0 Figure 8. Annual catches of swordfish reported in the Mediterranean Sea (Dr P. Kebe, 12 May 1999). Other fishing gear includes catches by driftnet, harpoon and unclassified gear (including some longline). observer data for Italy s longliners operating in the Ligurian Sea show that nearly 75 per cent of the longline catch comprises species that are commercially valuable (Orsi Relini et al. in press a). Swordfish make up 65 per cent of the catch, with northern bluefin tuna comprising 7 per cent and blue shark 3 per cent. Distribution Swordfish are found throughout the Mediterranean Sea, but are less abundant in eastern waters and in shallow waters fringing the coastlines of Tunisia (northern Africa). Most of the fishing for swordfish is in the Mediterranean s western basin and in the north-west of the eastern basin. The peak fishing season is July September, although large longliners operate usually year-round, following target species over a wide area. Swordfish fisheries 39

6 Italy s swordfish fleet operates in all of Italy s seas, except the North Adriatic Sea. The driftnet fishery occurs in the central and southern Mediterranean from April to August, whereas longliners operate in all seas, all year (GFCM ICCAT 1998, p. 14). Spain s swordfish fleet has traditionally fished in waters from Cabo de Creus, in the Strait of Gibraltar, across to the Balearic Islands, which is a swordfish spawning area (Folsom 1997b, p. 101). Most fishing activity occurs during June November (GFCM ICCAT 1998, p. 15). Many of the longliners then target northern bluefin tuna during May and June. Most of Morocco s swordfish fleet operate in Morocco s waters of the Strait of Gibraltar. They take their largest catches during September October, when swordfish are believed to migrate through the narrow strait (Folsom 1997a). Greece s swordfish fleet is concentrated mainly in the Dodecanese Islands, the Saronikos Gulf, the Peloponnesus and Crete (Folsom 1997b, p. 52). Most of Greece s longliners are based in Kalymnos and Chania. The longliners primarily fish the central and southern Aegean and Ionian Seas. The larger longliners are based in Crete and the Dodecanese Islands, because they fish in waters further from port, in the southern Ionian Sea and as far as Cyprus in the Levant Sea. The swordfish fishery operates mainly during May September. Poor weather and better market prices for other fish result in many of the longliners targeting other species for the remainder of the year. Cypriot fishers initially fished in their coastal waters along the southern, western and northwestern Cyprus coastline, although there are only limited productive fishing grounds within this area (Folsom 1997b, p. 137). Cyprus boats are also active in the eastern Mediterranean between Syria and Egypt. Tunisia s longliners operate mainly in the coastal waters off northern Tunisia in the Straits of Sicily. Turkey s fleet operates mainly in the Gulf of Anatalya (southern Turkey) and around the island of Gökçeada in the northern Aegean Sea (GFCM ICCAT 1998, p. 15). Turkish fishers also catch swordfish in the Sea of Marmara and the Black Sea (Folsom 1997b, p. 163). Japan s longliners operate seasonally, from April to July each year, in that part of the Mediterranean Sea west of 20ºE. Effort is most concentrated in waters around the Balearic Islands, Sicily, Tunisia and Libya (GFCM ICCAT 1998, p. 14). Oceanography A 400 m deep ridge between Sicily and the northern coastline of Africa divides the Mediterranean Sea into an eastern basin and a western basin (Pickard and Emery 1982, p ). The western basin reaches a depth of 3400 m in the central Tyrrhenian Sea. The eastern basin is deeper, reaching 4200 m in the southern Ionian Sea. The Mediterranean is characterised by high water temperatures and salinities compared to the major oceans. Because of the large heat input and much greater evaporation than rainfall, the temperature and salinity are higher in the eastern basin. Globally, the geographical distribution of swordfish is contained by the 13ºC sea surface isotherm. During January March, average monthly sea surface temperatures in the Mediterranean range from 13 to 14ºC in the western basin and 14 16ºC in the eastern basin. In July September, the range is 19 26ºC in the west and 20 30ºC in the east. Two main subsurface water masses characterise the Mediterranean Sea oceanography: The deep and bottom waters are formed at the northern edges of each basin, along the northern Ligurian Sea (western basin) and in the southern Adriatic Sea (eastern basin). The water is cooled by the strong northerly offshore winds that dominate the region in January March and mixes down to a depth of 1400 m. The water then rapidly sinks, to depths of more than 2500 m, inducing upwelling of colder, older deep waters. Formation of the deep water during January March is very rapid. The deep waters of the Mediterranean present an ideal habitat for swordfish. They are much warmer (12 13ºC) than those of the Atlantic (~2ºC) and contain much higher levels of dissolved oxygen ( mg/l). The Levantine Intermediate Water is formed during January March off the southern coastline of Turkey (Gulf of Anatalya Cyprus area). With a temperature of 15ºC and salinity concentration of 39.1 parts per thousand, the water flows westwards at a depth of m 40 Bureau of Rural Sciences

7 Mediterranean along Africa s northern coastline. It then flows out through the Strait of Gibraltar below a surface inflow of water from the North Atlantic. The water that emerges from the Strait of Gibraltar is the intermediate water and not water which has been displaced from the deeper basins. Gear and targeting Fleets The fleets fishing for swordfish in the Mediterranean consist of large, modern longliners and numerous small boats undertaking short (one or two-day) trips and using a variety of fishing gears (harpoon, driftnet and longline). Italy has the largest fleet of driftnetters targeting swordfish, although most of the fleet also fish for albacore outside the main swordfish season. Estimates of the driftnet fleet ranged from 472 to 682 boats in 1992 (Folsom 1997b, p. 66). The United Nations driftnet-length ban does not appear to have reduced the fleet s size in 1995 there were up to 680 licensed driftnetters, with 595 of those actively fishing. The driftnetters are relatively small (10 22 m, GRT). In 1992 Italy had a fleet of 701 longliners and 14 harpoon boats. All of the harpoon boats operate in the Straits of Messina. Spain s fleet of longliners fishing for swordfish in the Mediterranean decreased from 145 boats in 1985 to 75 boats in 1993, presumably because of declining catch rates (Folsom 1997b, p. 101). Most of the Mediterranean longliners are about 20 GRT or smaller. Few have refrigeration facilities and most return to port each day. Morocco s Mediterranean fleet consists of about 140 boats, mostly driftnetters that are based in Tangier (GFCM ICCAT 1998, p. 15). In 1997 the boats averaged 13 m and 50 GRT, although most of those fishing for swordfish are quite small (~10 GRT). Greece s swordfish fleet also consists of mostly small boats (<5 GRT). In 1987 the fleet consisted of about 334 boats; by 1994 it had increased to 400 boats. Many other small boats fish opportunistically. However, those have recently been limited by regulations introduced by the Ministry of Agriculture. Cyprus s swordfish fleet includes two types of boat: multipurpose boats and swordfish boats (Folsom 1997b, p. 138). The multipurpose boats are wooden and less than ten m in length. They carry a crew of three and return to port each day. These boats target swordfish only during the peak season (May June) and target demersal species for the remainder of the year. The swordfish boats are also wooden, but they are larger (10 15 m) and carry a crew of four. They are equipped with line haulers, radar and navigational aids. They also have freezers that allow the boats to remain at sea for up to ten days per trip. The boats target swordfish from March to October and may venture up to 275 nm from their home port. The swordfish fleet in Malta has steadily decreased in size from a peak in the early 1980s. Due to declining catches of swordfish, many boats switched to targeting more valuable northern bluefin tuna (GFCM ICCAT 1998, p. 14). In 1998 Malta s swordfish fleet consisted of ten longliners. The remainder of the fleet are multipurpose longliners that adapt their gear seasonally to target tuna and mahi mahi in addition to swordfish. During the peak swordfish season (June August) as many as 60 boats actually target swordfish. Larger boats operate 20 nm beyond the coastline, remaining at sea for at least five days per trip. The smaller boats fish closer to the coastline and stay at sea for less than four days. Tunisia s longline fleet consists of 40 boats between 9 and 24 m. The fishery targeted northern bluefin tuna in the past, but in recent years has increasingly targeted swordfish. Turkey s fishing fleet consists mostly of 1 5 GRT boats and few are equipped with motors or refrigeration equipment (Folsom 1997b, p. 164). There is no information on the number of boats in the swordfish fleet. Fish-finding We did not find any published information on techniques used to find swordfish by the Mediterranean fleets. Swordfish fisheries 41

8 Fishing gear and practices Originally, rope was used for the mainline, but nylon mainlines are now more common. The original technique employed by the Sicilians uses km long mainlines with branchlines every 35 m (Folsom 1997b, p. 52). The branchlines are usually 4 5 m long, with comparatively large (size 0 2) hooks. They usually use mackerel or sardine as bait. The gear is set at sunset, soaked overnight and retrieved at dawn. The same technique is probably common to most of the Mediterranean swordfish longliners, with smaller boats tending to set shorter longlines. The Italians also catch swordfish using driftnets and harpoons. Originally, they used driftnets that were about six km long, but some boats fished nets up to 20 km long (Folsom 1997b, p. 67). The nets are now made of nylon, with a mesh size of cm, although multistrand nets of 9 15 cm mesh size have been used in the Balearic Sea in the past. The nets have a deployed height of m. They are usually set perpendicular to the coastline, in a zigzag pattern. Driftnets are usually set late in the afternoon, taking three to four hours to complete. Retrieval of the net begins late at night and is usually completed by dawn. A 1995 study estimated that Italy s driftnet boats averaged 40 sets per year (Folsom 1997b, p. 67). Spanish fishers have used mostly pelagic longlines, called volantero, to catch swordfish. They adapted the gear from the traditional Sicilian longline technique. In the Mediterranean, the Spanish set about 1200 hooks per set on longlines that average 30 km in length (Folsom 1997b, p. 102). They attach branchlines about 30 m apart and there are usually six branchlines between each buoy. A radar reflector with light is attached every two km. They prefer mackerel as bait, although they sometimes use sardine, squid or shark. Spain s swordfish longliners set sail each afternoon and deploy their gear by dusk. The gear is left to fish through the night and is hauled before sunrise. They then return to port to market their catch in the early morning. Spain also catches swordfish using driftnets, but this fleet operates mostly in the Strait of Gibraltar. Since its fishery began, Greece s longliners have used traditional rope gear set to a depth of 30 m, with mackerel as bait, to catch swordfish. In recent years, the traditional gear has been gradually replaced by monofilament main and branchlines, which are set deeper (up to 60 m below the sea surface), with squid as bait and a lightstick attached (GFCM ICCAT 1998, p. 13). With remarkable similarity to the evolution of the Florida-style gear used in the north-west Atlantic, the distance between the branchlines has trebled and the number of hooks has been considerably reduced. The changes have been very successful, resulting in catches of bigger swordfish and increased bluefin bycatch. Cypriot fishers catch swordfish using pelagic longlines that are km long (Folsom 1997b, p. 138). The gear is set before sunset and hauled before dawn each day. The mainline is set very taught while the boat is at cruising speed and hooks are spaced about 20 m apart, with a buoy attached for every five hooks. Mackerel is the most common bait. Turkish fishers catch swordfish using longlines, harpoons and driftnets. The harpoon fishery occurs in the northern Aegean Sea, whereas longlines and driftnets are the primary gear in the Gulf of Anatalya. The longlines are usually about 12 km long with about 200 hooks. The driftnets are cm mesh size. The number of hooks deployed by Malta s longliners varies with boat size. Large longliners deploy up to 2000 hooks per set, whereas small longliners usually deploy hooks (GFCM ICCAT 1998, p. 14). Invariably the bait is Atlantic mackerel. The size of the bait used varies seasonally depending on the expected size of the swordfish. Malta s longliners, for example, use small mackerel when juvenile swordfish are expected. The Japanese use synthetic gear and set about 3000 hooks per day, with the hooks at m below the surface (Takeuchi 1996, p. 108). Squid baits are the most common. To catch northern bluefin tuna, the Japanese set their longlines an hour or two before dawn and haul during the late afternoon and evening. Discarding of target species There is a paucity of information on discarding practices of boats targeting swordfish in the Mediterranean, presumably because observers have rarely monitored fishing activities. 42 Bureau of Rural Sciences

9 Mediterranean Assessment Data collection Spain has been the nation most active in sampling swordfish catches in the Mediterranean, mostly as a result of its involvement with the International Commission for the Conservation of Atlantic Tunas (ICCAT) in the North Atlantic. The Instituto Español de Oceanografía (IEO) is responsible for monitoring Spanish catches. Since 1974 the Institute has collected data on Spain s swordfish landings through port sampling at the most important landing ports. However, no effort or catch rate information was collected in the Mediterranean until 1987 (Folsom 1997b, p. 108). Spain has occasionally placed observers on longliners, allowing the collection of sex composition data. During Italian scientists surveyed national catches (GFCM ICCAT 1996). In 1995 the National Institute of Statistics implemented a new sampling scheme. The scheme collected information from representative ports, without any statistical basis to the sampling. Italy intends to implement an observer program for their longline fishery by In Morocco, the L Institut National de Recherce Halieutique is responsible for monitoring swordfish landings. It conducts a size sampling program at the ports of Nador and Tangier (Folsom 1997a, p. 106). We were unable to determine the time frame or coverage of the sampling program. Fisheries research in Greece is undertaken by two government departments: the Fisheries Research Institution of the Ministry of Industry, Energy and Technology; and the Fisheries Laboratory of the Ministry of Agriculture (Folsom 1997b, p. 54). Greek scientists made an initial collection of catch and effort data from the pelagic fishing fleet in The Fisheries Department of Cyprus is responsible for monitoring Cyprus fishing fleet. The swordfish fleet has been monitored at varying levels since Research Research conducted on Mediterranean swordfish includes various studies of genetics, age and growth, reproduction and parasites. But, over most of the fishery s history, there has been no central agency undertaking biological research on Mediterranean swordfish. Instead, national governments and universities have undertaken a variety of research projects that have not always been directly relevant to assessment of the region-wide swordfish stock. During , however, Italian, Greek, Spanish and French scientists collaborated in a European Union (EU) project to identify the biological parameters essential to manage the fisheries for swordfish, northern bluefin tuna, albacore and Atlantic bonito. The project also attempted to obtain more realistic estimates of total landings, catch and size composition and catch rates for each of the fisheries than have been available in the past. The results of the project have significantly improved the basic data used by the General Fisheries Council for the Mediterranean ICCAT Working Group on Stocks of Large Pelagic Fishes in the Mediterranean Sea (the GFCM ICCAT Working Group ) to assess the Mediterranean swordfish stock. Stock assessment Fishery indicators Scientists have monitored the catch and effort of one of Italy s swordfish fleet, based in Port Cesareo in the northern Ionian Sea, every year since 1978 (except 1989; De Metrio et al. 1999). That data set has allowed comparison of indicator values for a discreet fleet operating in the same area each year. The average weight of swordfish landed by Cesareo longliners decreased continuously from 1978 to Longline catch rates also provide evidence of a decline in swordfish size. In terms of weight, catch rates declined during By comparison, catch rates in numbers of swordfish remained static, indicating that the longliners are still catching as many swordfish, but they are generally smaller than 20 years ago. Swordfish fisheries 43

10 Port Cesareo-based driftnetters operate in the same area as the longliners. Driftnet catch rates are confounded by under-reporting of net length by some Italian fishers. However, the average weight Year 0 Figure 9. Average weight of swordfish reported by Italy s Port Cesareo driftnet boats (De Metrio et al. 1999). of swordfish landed by the driftnet boats show an almost continual decline since 1980 (Figure 9). Nominal catch rates of Italy s longline fleet in the Ligurian Sea declined until 1992, when they reported their lowest catch rate (65 kg/1000 hooks; Orsi Relini et al. in press b; Figure 10). After 1992 catch rates gradually increased, reaching 118 kg/1000 hooks in The nominal catch rates of Japan s longliners operating in the Mediterranean has fluctuated. However, the overall trend has been declining since 1978 (0.28 swordfish/ 1000 hooks) to less than 0.05/1000 hooks in 1994 (Takeuchi 1996). The average weight of swordfish caught by Cyprus s swordfish fleet has declined from nearly 60 kg in 1976 to 20 kg in 1993 (Economou and Konteatis 1995, p. 344). The apparent decline in size might be related to an expansion in the area fished. However, declines in abundance cannot be ruled out. No size frequency data are available for the Cyprus fishery during that period (ICCAT 1995, p. 89). Few have attempted to standardise catch rates for Mediterranean swordfish fisheries. Spanish scientists have standardised their Mediterranean longline data, which covers nearly trips during , for year, area and quarter effects using a general liner model (GLM; Ortiz de Urbina et al. 1999). The catch rates in weight of swordfish showed no clear trend, whereas catch rates in terms of number of swordfish declined between 1988 and Stock assessment models A 1990 GFCM ICCAT Expert Consultation was the first major review of the data available for swordfish that encompassed the nine major nations fishing in the Mediterranean. No assessment was made. The meeting facilitated the collation of data on Mediterranean swordfish and defined methods for future data collection. 44 Bureau of Rural Sciences

11 Mediterranean In 1992 ICCAT s Standing Committee on Research and Statistics (SCRS) reviewed Mediterranean catch-at-size and catch rate data and decided that the application of stock assessment models was premature because of the short time-series of catch-at-size data and the lack of standardised catch rates (ICCAT 1993, p. SWO 21). SCRS did note a decline in the Year 0 Figure 10. Nominal catch rates reported by Italy s longliners in the Ligurian Sea (Orsi Relini et al. in press b). average size of Mediterranean swordfish, from 127 cm in 1985 to 115 cm in 1992, as estimated from the combined catch-at-size data. Recognising that the Mediterranean catch levels were comparable to those of the entire North Atlantic, SCRS stressed the need for a proper stock assessment and encouraged scientists from all GFCM member nations to participate, as only Spain had so far. In 1995 a GFCM ICCAT Working Group developed a tuned virtual population analysis (VPA) for Mediterranean swordfish. They used a sex-specific, catch-at-age analysis, tuned with several standardised catch rate series. They derived the catch rates from data reported by longliners of Italy, Greece and Spain and Italy s driftnetters. 2 The analysis indicated stable or increasing abundance between 1985 and 1994, with increased fishing mortalities in the late 1980s. Fishing mortalities then decreased until 1994 (GFCM ICCAT 1996, p. 14). The Working Group used yield per recruit and spawning-biomass per recruit analyses to examine how changes in fishing mortality would affect annual recruitment (GFCM ICCAT 1996, p. 15). While the actual fishing mortalities of recent years were very uncertain, the calculations suggested that a per cent increase in yield per recruit could be gained by reducing the catch of cm swordfish. The spawning potential of the stock in 1994 was estimated to be per cent of that for the virgin stock. Without any knowledge of the rate of recruitment into the fishery, the high catches of juvenile swordfish and the scarcity of large swordfish in many of the catches, are cause for concern. The fishery is based on only two or three, mostly immature, age classes and it is vulnerable to changes in recruitment. The age and size at maturity of Mediterranean swordfish is substantially less than that observed for Atlantic swordfish, which may be a result of either prolonged heavy fishing pressure or different environmental conditions in the Mediterranean. Alternatively the small age 2 The analysis used catch and effort data from Greece s longline fishery for (quarterly catch and effort) and (monthly); from Italy s longline and driftnet fisheries (except 1993); and from Spain s longline fleet (GFCM ICCAT 1996, p. 13). Swordfish fisheries 45

12 and size at maturity might be a characteristic of the Mediterranean stock and explain the high productivity of swordfish there. The results of other biological research support the conclusion that Mediterranean swordfish are overexploited. An analysis of genotype variability in Mediterranean swordfish revealed that their genetic diversity is low when compared to Atlantic swordfish caught off South Africa, where fishing pressure has been comparatively low (GFCM ICCAT 1998). Fish populations often become prone to parasites and disease when their genetic diversity is reduced. Mediterranean swordfish have been found to have a higher prevalence of parasitical nematodes and copepods than swordfish outside the Mediterranean. This supports the hypothesis of genetic erosion resulting from prolonged high levels of fishing mortality (GFCM ICCAT 1998, p. 17). However, alternative explanations of the heavy parasite loads are plausible, e.g., the warm conditions or the high density of swordfish in the Mediterranean might be conducive to parasites. Status In view of the short time-span of reliable data and the extraordinarily high catches of juvenile swordfish, in 1995 the GFCM ICCAT Working Group suggested that it was unlikely that the Mediterranean Sea could continue to yield such high catches of juveniles without continued high recruitment. They also noted that the probability of high recruitment will diminish as mature swordfish are removed from the population. Consequently, they recommended a substantial reduction in fishing mortality, particularly on pre-spawning juveniles the highest priority until the status of the stock can be determined with more certainty. Reliability of the assessment There is concern over the reliability of the assessment because of recent revisions and adjustments made to the Mediterranean catch data sets (GFCM ICCAT 1998, p. 16). Many nations have made no report, while others have revised only parts of the time series of data. For the 1995 assessment, the GFCM ICCAT Working Group created a sex-specific catch-at-age database mostly from the data of Spain, Italy and Greece, with many substitutions of data between fishing fleets and areas to complete the series. The sex-ratios used to convert the data to a sex-specific series were limited to Spain s longline data. 3 However, the sex-ratio of swordfish in those data varied between coastal and offshore areas and between fishing gears. Consequently, the application of Spain s sex-ratios to a Mediterranean-wide assessment might not accurately portray the true situation (Mejuto et al. 1994). The Working Group recommended that a representative sample of the catch be sexed when measured at sea, by fine area and time strata for each gear and fleet (GFCM ICCAT 1996, p. 11). In 1998 the GFCM ICCAT Working Group pointed out that the reported catches of Mediterranean swordfish are likely to be underestimated and the true catch level is not really known. When swordfish are caught as a bycatch of another fishery, they are often not reported. Information is not available on the many landings made by boats in the numerous small European harbours. Swordfish are landed by Albania, Lebanon, Israel and, possibly, Egypt. However, no data exist for those fisheries. The catch of Cyprus is likely to have been significantly underreported, whereas the catch of Tunisia was revised by such a large amount that the Working Group expressed considerable concern over the magnitude of non-reported catches (GFCM ICCAT 1998, p. 19). Morocco recently revised the estimates of its catches by more than 100 per cent. 4 A large portion of Morocco s catch comes from the Strait of Gibraltar and the catch was previously classified as taken in the Atlantic. While it is now generally accepted that the Atlantic and Mediterranean swordfish constitute separate stocks, it is not clear which stock the swordfish in the Strait belong to and both Morocco and Spain take appreciable catches in the area. In 1998 Morocco s catches from the Strait were moved from Atlantic to Mediterranean catches. 3 Nearly samples were averaged by quarter with all years combined to give quarterly sex ratios (GFCM ICCAT 1996, p. 11). 4 The catch of 1995 was revised upwards by nearly 300 per cent (GFCM ICCAT 1998). 46 Bureau of Rural Sciences

13 Mediterranean However, the driftnet catches made by Spain in the same area were not moved, leading to further inconsistencies in the data. The magnitude of the changes caused to both North Atlantic and Mediterranean assessments is likely to be significant because of the large amounts of swordfish taken in the Strait. In 1998 the GFCM ICCAT Working Group highlighted the need for all nations currently fishing for swordfish to collect catch and effort data in time and area strata that are small enough to allow fine-scale standardisation of catch rates. They recommended that longline data be recorded for individual sets, with precise set position recorded, rather than total catch for entire trips. Many nations will need to implement comprehensive logbook programs and appropriate measures to ensure compliance and reliability of the data. The Working Group also recommended that scientists undertake appropriate size and sex sampling by area each month. Such sampling would require observer programs because all swordfish are landed in a gilled and gutted condition in the Mediterranean. Studies on the selectivity of different gears for catching swordfish are needed to develop appropriate models for standardising catch rates. The effects of sex, changes in fishing strategies and environmental conditions also need to be included in future analyses. In addition to the problems with estimates of catch levels and composition, uncertainties in biological parameters (particularly a validated age key) natural mortality and sex ratio, hamper the virtual population analyses. Continued catches of small swordfish suggest that recruitment can be maintained at very low levels of parent stock, but many of the life-history parameters of swordfish in the Mediterranean are poorly known (GFCM ICCAT 1998, p. 19). A better understanding of fecundity, spawning and the distribution of eggs and larvae is needed for swordfish in all areas of the Mediterranean. Tag recapture studies, particularly the use of archival tags, could be used to determine daily and seasonal patterns of migration related to feeding and reproductive behaviour. The origin(s) of the stock being fished in the Strait of Gibraltar needs to be determined and rules need to be established for allocating those catches to the Mediterranean or Atlantic stocks. In the interim, the GFCM ICCAT Working Group recommended that stock assessments in each area consider the impact of the uncertainty in the origin of the catches through sensitivity analyses (GFCM ICCAT 1998, p. 18). Management Institutions Nations independently manage their Mediterranean swordfish fisheries. GFCM provides loose coordination of national management measures and promotes data exchange. It has not systematically reviewed the extent to which its members have adopted its recommendations (GFCM 1997, p. 3). European Union members also adopt recommendations agreed to by the European Council (EC). Management measures European Union members enforce the European Council s minimum size limit of 120 cm for Mediterranean swordfish (GFCM ICCAT 1998). Some non-members, such as Turkey and Croatia, also observe the size limit. Both GFCM and European Union members observe the United Nations-initiated ban that limits pelagic driftnets to 2.5 km in length. The European Council regulation also requires that all driftnets remain permanently attached to the boat. In general, however, most fleets using driftnets in the Mediterranean do not comply with the regulations and regularly deployed pelagic driftnets longer than 2.5 km until at least 1994 (GFCM ICCAT 1995, p. 10). Since the early 1980s Italy has imposed a minimum size regulation that permits no more than ten per cent of the swordfish catch to have a length less than 140 cm (ICCAT 1993, p. SWO 21). But it is unclear how Italy enforces the regulation. As early as 1991 Italy had introduced regulations that limited driftnets to 2.5 km. Yet, there are many anecdotal reports of Italian fishers continuing to use much longer driftnets well after the 1992 United Nations ban. Associated incorrect reporting of gear dimensions and types also creates problems for standardising Italy s catch rates. Consequently, any trends in driftnet catch rates are unlikely to accurately reflect Swordfish fisheries 47

14 swordfish abundance. In 1998 Italy began to reduce its reported driftnet fishery, mostly through enforcement of net length restrictions. They also banned the use of driftnets in the Ligurian Sea to protect large numbers of cetacean there (GFCM ICCAT 1998, pp. 14, 19). Morocco manages its swordfish fishery by a combination of limited entry, effort limits, time and area closures and size limits (Folsom 1997a, p. 106). Morocco imposes a minimum size limit of 125 cm for swordfish. The length of driftnets is limited to 2.5 km with a minimum mesh size of 400 mm. Only one longline may be carried per boat. Only swordfish longliners of greater than 15 hp capacity may use driftnets and no new boat may enter the fishery. Fishing is only allowed during December June between Nador and Ras Kebdana and during April October from Al Hoceima to Tangier. Fishing gear may only be set in a six hour period each day commencing at sunset. Greece has banned fishing in its seas during October January each year to protect very small (age 0) swordfish (GFCM ICCAT 1998, p. 13). But, Greek biologists claim the ban has had little effect because: it does not stop other nations from fishing the same juvenile swordfish in adjacent waters (Greece s territorial waters are less than 6 nm wide); many small juveniles are still landed during February, indicating that the fishery is reopened too soon; and it does not stop boats landing swordfish that they claim were caught as bycatch (Folsom 1997b, p. 54). In addition to enforcing the European Council s size limit of 120 cm, Spain has restricted its longliners in the Mediterranean to 2000 hooks per set (GFCM ICCAT 1998, p. 19). Turkey implements the minimum size limit and has a closed season during July September. Cyprus also has a closed season for swordfish, from October to February. In 1990 Cyprus introduced licences for swordfish boats, limiting the number of boats to 60 each year (Economou and Konteatis 1995). To protect spawning northern bluefin, Japan prohibited its longliners from fishing in the Mediterranean between 21 May and 30 June during Since 1995 Japan has observed the GFCM recommendation to disallow fishing by longliners longer than 24 m during June July. Japan has voluntarily limited its fleet to a maximum of 35 longliners since 1985 (Takeuchi 1996, p. 108). The effectiveness of minimum size regulations for Mediterranean swordfish is doubtful because of the high proportion of juvenile swordfish in current catches and the expected low survival rates of swordfish released after being captured by longline. At present it is not clear how nations actually enforce the size rule. Consequently, the GFCM ICCAT Working Group proposed the following alternative management measures for Mediterranean swordfish: 1. the present catch levels for swordfish be maintained or reduced through appropriate arrangements, possibly quotas; 2. local closed areas or seasons be implemented to protect spawning and juvenile swordfish; or 3. improvements in gear selectivity be developed (GFCM ICCAT 1998, p. 19). Bycatch There have been few independent estimates of bycatch in the Mediterranean swordfish fisheries. Observer data from Italy s longliners in the Ligurian Sea show that swordfish made up 65 per cent of the catch, with the remainder consisting of northern bluefin tuna (7%), blue shark (3%), Ray s bream (3%), mahi mahi (2%), wreckfish (2%), sandbar shark (<1%) and Atlantic bonito (<1%; Orsi Relini et al. in press a). The observer data indicate that loggerhead turtle, manta ray and ocean sunfish are occasionally caught by longliners. They are usually released alive by cutting the branchline (Orsi Relini et al. in press a). Marine mammals incidentally caught by longlines in the Mediterranean include striped 48 Bureau of Rural Sciences

15 Mediterranean dolphin, Risso s dolphin, bottlenose dolphin, false killer whale, fin whale and monk seal (Di Natale 1991). Concern over the bycatch of cetacean by driftnets operating in the Ligurian Sea led to a 1990 agreement between France, Italy and Monaco to ban the use of driftnets in a 6000 nm 2 area of the central western part of the Ligurian Sea (GFCM ICCAT 1995, p. 9). Interaction Recreational anglers rarely target swordfish in the Mediterranean. However, there is intense interaction between the various fishing gears (e.g., driftnet and longline) and the many fleets fishing for swordfish. Several nations take swordfish as a bycatch of northern bluefin-directed fishing operations and northern bluefin are taken as a bycatch of swordfish longlining. The amounts and types of that bycatch are poorly reported in most Mediterranean fisheries, so the magnitude of the catches is not well known. Advice needs The lack of accurate estimates of current and historical catch levels is a fundamental problem hampering management of the Mediterranean swordfish fishery. The fishery cannot be managed effectively without detailed and comprehensive knowledge of current and historical catch levels. It will take a major, concerted effort over many years to assemble a reasonably accurate time series of data suitable for a reliable assessment of the Mediterranean swordfish stock. Further yield per recruit analyses might help to demonstrate the increased harvests (and financial gains) that a reduction in juvenile mortality might create. Fishery managers, industry and scientists also need to explore options for reducing the mortality of juvenile swordfish. Size restrictions are believed to be ineffective because of high mortality rates of released swordfish, but this needs to be verified and quantified. Dedicated research might reveal practical modifications to fishing gear that will reduce the catch of small swordfish. Area and season closures are other options that need investigation through analyses of fine-scale data. Uncertainty over mixing of swordfish between the Atlantic and Mediterranean through the Strait of Gibraltar has the potential to create regulatory loopholes. It also contributes to uncertainty in stock assessments. A tag recapture experiment would be useful in estimating mixing rates through the Strait. Such an experiment might also provide useful estimates of population size and population parameters, such as natural mortality, fishing mortality and growth. However, there are practical difficulties in tagging the numbers of swordfish that such an experiment requires. Furthermore, analyses of tag recapture data would require accurate estimates of reporting rates and detailed catch, effort and size data for the fisheries. Regardless, the lack of a regional management body with powers to enforce regulations throughout the Mediterranean presents a major obstacle to rebuilding the swordfish stock and improving the utilisation of the resource. Future prospects The uncertain status of the stock and the fishery s heavy reliance on immature swordfish are causes for concern. Continued mortality of juvenile swordfish may cause significant reductions in the size of the parent stock and wide fluctuations in recruitment. A fishery collapse may result with little warning. Yet, it is unlikely that the diverse nations fishing in the Mediterranean will be able to agree on and enforce appropriate management measures for swordfish. Consequently catch levels and returns from the resource are likely to stay well below optimum. References Buencuerpo, V., Rios, S and Moron, J. (1998) Pelagic sharks associated with the swordfish, Xiphias gladius, fishery in the eastern North Atlantic Ocean and the Strait of Gibraltar. Fishery Bulletin 96, pp De Metrio, G. (1995) Characterisation of large pelagic stocks in the Mediterranean European Union Project XIV/MED/91 102: Description of the project and synthesis of the research on swordfish. Collected volume of scientific papers 44(1), pp Swordfish fisheries 49

16 De Metrio, G., Cacucci, M., Megalofonou, P., Santamaria, N. and Sion, L. (1999) Trends of swordfish fishery in a northern Ionian Port in the years between 1978 and Collected volume of scientific papers (in press). ICCAT, Madrid. Di Natale, A. (1991) Interaction between marine mammals and scombridae fishery activities: The Mediterranean case. In GFCM ICCAT. Report of the GFCM ICCAT Expert Consultation on Evaluation of Stocks of Large Pelagic Fishes in the Mediterranean Area, Bari, Italy, June Food and Agriculture Organization Fisheries Report No Economou, E. and Konteatis, D. (1995) Review of swordfish fishing pp In Collected volume of scientific papers 44(1). ICCAT, Madrid. Folsom, W.B. (1997a) World swordfish fisheries: An analysis of swordfish fisheries, market trends and trade pattern, past present future, Volume II Africa and the Middle East. United States Department of Commerce, NOAA Technical Memorandum NMFS F/SPO 24. Folsom, W.B. (1997b) World swordfish fisheries: An analysis of swordfish fisheries, market trends and trade pattern, past-present-future, Volume VI Western Europe. United States Department of Commerce, NOAA Technical Memorandum NMFS F/SPO 24. GFCM (1997) A medium and long term work programme for GFCM; implication for institutional structure and resource needs. Document GFCM/XXII/97/3 presented at the Twenty-second Session of the General Fisheries Council for the Mediterranean, Rome, October Food and Agriculture Organization, Rome. GFCM ICCAT (1995) Report of the Ad Hoc GFCM ICCAT Joint Working Group on Stocks of Large Pelagic Fishes in the Mediterranean Sea, Fuengirola, Malaga, Spain, September 19 to pp of Collected volume of scientific papers 44(1). ICCAT, Madrid. GFCM ICCAT (1996) Report of the Second Meeting of the Ad Hoc GFCM ICCAT Joint Working Group on Stocks of Large Pelagic Fishes in the Mediterranean Sea, Bari, Italy, September 13 to Mediterranean Swordfish Data Preparatory Meeting. pp of Collected volume of scientific papers 45(1). ICCAT, Madrid. GFCM ICCAT (1998) Report of the Fourth Meeting of the Ad Hoc GFCM ICCAT Joint Working Group on Stocks of Large Pelagic Fishes in the Mediterranean Sea, Genoa, Italy September 7 to ICCAT, Madrid. ICCAT (1993) Swordfish species section 1992 final SCRS report. Collected volume of scientific papers 40(1). ICCAT, Madrid. ICCAT (1995) 1994 SCRS report Mediterranean swordfish. Report for the biennial period , part 1 (1994), Vol 2. ICCAT, Madrid. ICCAT (in press) 1998 SCRS report Mediterranean swordfish. ICCAT, Madrid. Megalofonou, P., De Metrio, G. and Lenti, M. (1990) Catch, size distribution, age and some population parameters of swordfish, Xiphias gladius L., in the Greek Seas. pp of Collected volume of scientific papers 33. ICCAT, Madrid. Mejuto, J., De la Serna, J.M. and García, B. (1994) An overview of the sex-ratio at size of the swordfish (Xiphias gladius L.) in the Atlantic and Mediterranean Sea: Similarity between different strata. pp of Collected volume of scientific papers 44(3). ICCAT, Madrid. Orsi Relini, L. Palandri, G., Garibaldi, F. and Cima, C. (in press a) Longline swordfish fishery in the Ligurian Sea: Eight years of observations on target and bycatch species. Collected volume of scientific papers. ICCAT, Madrid. Ortiz de Urbina, J., de la Serna, J.M. and Mejuto, J. (in press b) Updated standardised catch rates in number and weight for the swordfish (Xiphias gladius) from the Spanish longline fleet in the Mediterranean Sea, Collected volume of scientific papers. ICCAT, Madrid. 50 Bureau of Rural Sciences

17 Mediterranean Pickard, G.L. and Emery, W.J. (1982) Descriptive physical oceanography: An introduction (4th edition). Pergamon Press, Oxford. Takeuchi, Y. (1996) Status report of the Japanese longline fishery in the Mediterranean sea with special reference to swordfish. pp of Collected volume of scientific papers 45(1). ICCAT, Madrid. Swordfish fisheries 51

18 The North Atlantic swordfish fishery Overview At least 15 nations have fished for swordfish in the North Atlantic, with Spain, the United States, Canada, Portugal and Japan taking the largest catches. Catches peaked at t in 1987 then declined due to the relocation of longliners to other areas, the introduction of quotas and swordfish size limits and reduced swordfish abundance. Since 1988 the International Commission for the Conservation of Atlantic Tunas (ICCAT) 1 has assessed the status of North Atlantic swordfish. In 1990 assessments by ICCAT s Standing Committee on Research and Statistics (SCRS) first showed a continuous decline in swordfish biomass since about 1980 with the stock eventually falling below the optimum biomass. Catch rates generally declined from 1990 to 1996 and there were fewer older-age swordfish in the population. But, the latest assessment (1999) provides a more optimistic outlook for North Atlantic swordfish. It suggests that the decline in swordfish biomass has slowed and that there were strong recruitments of young (age 1) swordfish in 1997 and Those swordfish should eventually contribute to the parent stock if they are not heavily harvested before they reach maturity. In response to the 1990 assessment, ICCAT issued its first recommendation calling for catch reductions. ICCAT subsequently recommended specific output control measures to manage the stock, e.g., total allowable catches (TACs) and minimum size limits. However, some member nations have ignored those recommendations while other measures (e.g., minimum size limits) have not been particularly effective. In response to declining catch rates and established international quotas, some longliners relocated to the South Atlantic, while others opportunistically target tuna and shark to take advantage of market prices and locally high catch rates. The fishery Development North-western Atlantic Swordfish have a long history of commercial exploitation in the North Atlantic. As early as 1840 a well-established coastal fishery off New England, United States, was landing between 3000 and 6000 swordfish per year. At an average weight of 114 kg, North Atlantic swordfish were harvested at t per year (Goode 1883, Cited in Hoey et al. 1989). By the late 1860s harpooning for swordfish was a widespread but opportunistic activity. Some of the groundfish fleet rigged lookout towers and carried harpoons to take swordfish when cod, haddock and flounder catches were poor or when they came upon large concentrations of swordfish (Dr J. Hoey, 26 February 2000). Prior to 1884 most of the United States swordfish catches were taken inshore, in waters around New Bedford, Martha s Vineyard and the Gulf of Maine (Gibson 1998; Map 3). Boats moved offshore in the late 1890s when swordfish became less abundant on the original inshore grounds. In 1903 a harpoon fishery began in Nova Scotia (Canada) and by 1909 it was important enough to be considered separately in the fishery statistics. By 1940 Canada was landing more than 80 per cent of the swordfish catch of 2.5 million pounds (~1134 t) in the north-western Atlantic (Tibbo et al. 1961).Until 1962 the north-western Atlantic fishery was solely a harpoon fishery targeting large, female swordfish that averaged around 120 kg, with swordfish smaller than 60 kg rare (Berkeley 1989, p. 51). Fishing occurred in waters north of 40ºN during July October. The swordfish resource seemed large and inexhaustible. 1 We use the acronym ICCAT to refer to deliberations of annual meetings of member nations under the ICCAT Commission. ICCAT s Standing Committee on Research and Statistics (SCRS) reviews the work of separate species working groups. SCRS then reports its conclusions to annual meetings of ICCAT members. For brevity we refer to SCRS as the source of assessments, whereas those assessments are a product of both SCRS and the Swordfish ( SWO ) Working Group. 52 Bureau of Rural Sciences

19 North Atlantic -50 Canada United States Cape Hatteras Gulf of Mexico Mexico Nova Scotia Georges Martha s Vineyard Bank Gulf Stream Bermuda Antilles Current Newfoundland Grand Banks Labrador Current 2000 m Straits of Florida Puerto Rico Sargasso Sea Guiana Current Mid- Atlantic Ridge Mid- Atlantic Ridge Azores Cape Verde Madeira Islands Canary Islands France Vigo Spain Portugal Venezuela North Equatorial Current Equatorial Counter Current Gulf of Guinea -0 Map 3. The northern Atlantic Ocean, showing approximate location of 200 nm exclusive economic zones, major ocean circulation systems (Pickard and Emery 1982, p. 139) and the 2000 m isobath. In the early 1960s incidental catches of swordfish by both Japan s and Norway s longliners prompted Canada s and United States boats to use pelagic longline gear to target swordfish. Experimental longlining produced some phenomenal catches (Hoey et al. 1989). Longline gear proved to be far more efficient than harpooning for catching swordfish. Not only were daily catches much higher, but boats could operate during inclement weather unsuitable for harpooning and could operate further offshore when waters cooled over the continental shelf. The result was an extension of the fishing season and expansion of the fishing area. Catch rates increased, but the average weight of swordfish declined because longline caught smaller male swordfish as well as female swordfish (Beckett 1971; Caddy 1976; Hurley and Iles 1981). 2 The longliners also fished in warmer waters where smaller swordfish were more abundant (Dr J. Hoey, 26 February 2000). In the late 1960s boats from New England experimented with longlining for swordfish in the Straits of Florida. Their early attempts, using shallow-set New England style gear in the 2 4 knot Florida Current, were mostly unsuccessful (Berkeley et al. 1981). As a result, boats moved to the Gulf of Mexico where they began to longline during January March when catch rates declined in the north. The Gulf of Mexico fishery had only just become commercially feasible when, in 1971, the United States Food and Drug Administration (FDA) established a 0.5 ppm tolerance level for mercury in swordfish a level that most swordfish exceed. The Food and Drug Administration ordered a ban in the United States on the sale of all swordfish exceeding the limit, while Canada officially closed its fishery. In the United States the market ban was not stringently enforced in 2 Cited by Dr J. Porter (16 December 1999). Swordfish fisheries 53

20 0 1 2 Landings 3 ('000 t) 4 5 Sp ain United States Jap an Canada Portugal Taiw an Morocco Venezuela Korea Mexico Others Figure 11. Annual landings of swordfish by the ten leading harvesters in the North Atlantic, 1997 (ICCAT database). several states. In consequence, swordfish fishing continued with United States catches not reported. Canada s boats that continued to fish illegally during the closure transferred their catch at sea to United States longliners (Gibson 1998, pp ). As a result of a successful legal challenge by the commercial fishing industry, the Food and Drug Administration revised the tolerance level for mercury to 1.0 ppm in The more liberal tolerance level allowed the United States fishery to legally continue. Canada officially reopened its fishery in 1979, but was unable to resume its dominance of the fishery. Significant gear modifications and improvements in technique during the late 1970s had allowed United States longliners to expand their operations, particularly in southern waters. In the mid-1970s professional charter boat operators began to use lightsticks while drifting at night to target swordfish in waters off south-east Florida (Dr J. Hoey, 26 February 2000). Cuban- American lobster fishers, displaced by the closure of shelf waters around the Bahamas, soon began experimental longlining in that area using techniques derived from the traditional Cuban longline fishery (Berkeley et al. 1981). Using shorter mainlines than the New England longliners, they were much more successful in the swift Gulf Stream waters. Their success, combined with the displacement of boats from newly declared territorial waters around Canada and Mexico, encouraged more longliners to move to Florida. The Florida-based fleet increased to more than 200 longliners by Many of the new entrants were small fibreglass boats crewed by Florida fishers formerly involved in charter fishing (Dr J. Hoey, 26 February 2000). United States longliners vary the amount of effort directed at swordfish with lunar phase and also seasonally. In 1987, for instance, nearly 70 per cent of United States longline sets targeting swordfish occurred during the two weeks around the full moons (Podesta et al. 1993, p. 259). The level of effort directed to swordfish was greatest during July and August and decreased after August when some swordfish longliners targeted tuna. The 1980s saw major improvements in gear efficiency and further expansion of the fishing grounds. In early 1985 swordfish longlining commenced in the Caribbean, when 13 United States longliners began fishing distant tropical waters north-east of Puerto Rico (Bannerot and Austin 1989, p. 182). The following year the fleet based in San Juan, Puerto Rico, swelled to 45, while others were fishing up to 1000 nm east of the Lesser Antilles (Gibson 1998, p. 42). From 1985 to 1989 the north-western fleets began fishing further east, around frontal zones north of the Azores. 54 Bureau of Rural Sciences

21 North Atlantic By 1989, 28 per cent of the United States Atlantic catch came from that region (Hoey et al. 1989). In the early 1980s a driftnet fishery for swordfish also began off New England (Brewster-Geisz 1997, p. 66). The number of United States boats longlining for swordfish in the North Atlantic declined every month during The decline was partly due to the movement of boats, particularly from the Gulf of Mexico, into the Hawaii-based fishery (Hoey et al. 1995; see p. 111). In Canada some swordfish longliners began to target tuna during (Stone and Porter 1998, p. 234). The shift was a direct result of declining swordfish abundance and reduced quotas for North Atlantic swordfish. In 1998 a coalition of environmental groups attempted to reduce pressure on the North Atlantic swordfish stock through the Give swordfish a break campaign. The campaign was a response to what the groups considered to be government inaction on scientific advice recommending reduced catches of swordfish. The resulting consumer boycott of swordfish in United States restaurants depressed swordfish prices in the United States (NRDC 1998, p. 3). Directed swordfish fisheries exist in several Caribbean nations, including Trinidad and Tobago, Grenada, St Vincent and Bermuda and also Venezuela and Mexico. Of those, Venezuela takes the largest catch (85 t in 1997) using longlines in its exclusive economic zone (EEZ; Weidner et al. 1999, p. 31). North-eastern Atlantic Longlining for swordfish began in coastal areas off south-western Spain in the late 1800s. The fishery, based in Algeciras, landed less than 1000 t of swordfish per year (Berkeley 1989, p. 48). The fleet rapidly expanded during 1966 when a new longline fishery developed off north-western Spain, from the port of Vigo. Landings trebled, reaching more than 3000 t a year and fluctuated around that level until 1980 when another period of expansion began. Some of the Algeciras longliners, which had been fishing around the Iberian Peninsula, began extending their operations off the coastline of Africa and into the Gulf of Guinea as far south as the equator. At the same time, Vigo longliners began moving westwards, eventually reaching the Azores. The northwestern fleet expanded the fishing grounds as far as 40ºW by 1983 and 45ºW by 1985; the fleet now fishes as far west as 50 W (Folsom 1997, p. 100). A gradual but continual decline in the average size of swordfish caught by the Spanish since 1975 (e.g., 88 kg on average in 1975 down to 56 kg in 1985) was the main reason for the expansion of fishing effort. On average, swordfish caught by longliners near the Canary Islands were smaller (44 kg, on average) than those caught further north, but catch rates (1.3 t/1000 hooks) were about four times higher (Berkeley 1989, pp ). Following further declines in catch rates in the north-eastern Atlantic, many of Spain s longliners moved to waters south of 5ºN during (Folsom 1997, p. 17). The Portuguese have caught small numbers of swordfish in waters off Portugal for possibly hundreds of years (Folsom 1997, p. 311). In 1965 they landed 6 t, caught mainly by harpooning or as a bycatch of trawling. The directed swordfish fishery began in 1984 when several local fishers purchased longliners. They were immediately successful and the landings increased from 22 t in 1984 to 468 t in 1986 (Folsom 1997, p. 86). Portugal s longliners also fish around the Azores and the Madeira Islands. Other nations involved in the North Atlantic swordfish fishery include the Azores, Morocco and Liberia. Several other west African states also catch small quantities of swordfish which are not reported. Distant-water fleets Japan, Taiwan and the Republic of Korea, have taken swordfish as an incidental catch to tuna longlining in the North Atlantic. Japan s longliners began fishing in the Atlantic in 1952 using rope and basket gear. They initially targeted yellowfin and albacore in tropical equatorial waters, moving further north when bigeye and bluefin became the target species (Uosaki and Uozumi 1993, p. 371). Although Japan s longliners do not target swordfish, in 1997 they landed more than 11 per cent of the total North Atlantic swordfish catch. For the entire Atlantic Ocean, Japan is the second largest harvester of swordfish (Dr J. Hoey, 26 February 2000). Swordfish fisheries 55

22 Taiwan and Korea commenced longlining in the North Atlantic in the early 1960s. Most of the activity of Taiwan s and Korea s longliners has been in the Indian and Pacific oceans and only minor catches of swordfish, incidental to their albacore and yellowfin tuna operations, are reported from the Atlantic. Catches Catch trends As a result of the introduction of longlining, North Atlantic swordfish catches increased from about t per year in the 1950s to in Catches then declined to around 9000 t per year for the remainder of the decade (Figure 12). The United States Food and Drug Administration mercury ban, in force from , resulted in reduced swordfish landings (SAFMC 1985, p. 17). Longline Other gear 20 MSY Catch ('000 t) Year 0 Figure 12. Annual catches of swordfish reported in the North Atlantic (ICCAT 1997c). Harpoon accounts for most of the catch classified as Other gear. Also shown is the maximum sustainable yield (MSY) estimated in SCRS s 1999 stock assessment. North Atlantic swordfish catches increased substantially to t in 1978, immediately following the revision of the Food and Drug Administration mercury standards. The catches then fluctuated between and t per year until 1986, when they rose to t. The boom was to continue for another year, when the peak North Atlantic catch of t was landed. The United States and Spain accounted for nearly 80 per cent of the landings, with increased harvests by both nations in the central North Atlantic responsible for most of the gain in landings (Hoey and Mejuto 1991, p. 416). ICCAT recommended international quotas for North Atlantic swordfish in 1992 (Dr J. Hoey, 26 February 2000). Anticipating the quotas, or in response to them, many longliners moved to other areas, e.g., Spain s and United States longliners to the South Atlantic, the Pacific and, more recently, the Indian Ocean (Porter 1997, p. 23). The quotas also affected fleets that could not relocate (Dr J. Hoey, 26 February 2000). During the total reported catch was around t per year. Total catch (landings plus discards) then declined to t in 1997 and t in 1998 (ICCAT in press b; ICCAT 1999b, p. 1).The decline was mainly the result of Spain and the United States halving their catches from the record levels of 1987 in response to ICCAT recommendations (ICCAT in press b, p. 48; NOAA 1997). Changes in fishing operations also contributed to reduced landings. Longliners of several fleets (e.g., Spain, Portugal, Canada and the United States) now opportunistically target tuna or shark to take advantage of market conditions and high catch rates of those species in some areas (ICCAT 1999c, p. 1). 56 Bureau of Rural Sciences

23 North Atlantic Size composition The average weight of swordfish landed in the North Atlantic has declined over the history of the fishery. Prior to 1963 swordfish averaged about 90 kg, compared to 45 kg in 1970 (SAFMC 1985). The declines in size can be partly explained by the change from harpoon fishing gear, which specifically targets large female swordfish, to longline, which typically takes a wider size range of males and females. The extension of fishing into tropical and equatorial waters, where smaller swordfish are more abundant and the season is longer, also contributed to the reduction in average size. Furthermore, markets influence the size of swordfish targeted, with markets in the United States and Canada often not accepting small (<15 kg T wt or ~20 kg W wt ) swordfish. Using comparative analyses based on nominal catch rate data by sex and area, Mejuto et al. (1998) and Ortiz et al. (2000) describe the patterns of sex-ratio at size in relation to the biology and behaviour of swordfish in the Atlantic. Such analyses lead SCRS to consider three types of behavioural patterns in the Atlantic: feeding, spawning and transitional. For example, longline catches tropical areas of the western Atlantic are characterised by a concentration of males and a paucity of females. The pattern of sex-ratio at size can be considered an indicator of important reproductive activity (Dr J. Porter, 21 June 2000). Since 1992 United States longliners have landed swordfish that range from 15 to 90 kg T wt (Cramer and Adams 1999). Many smaller swordfish are caught. Since 1991 the United States has enforced the ICCAT recommendation for a minimum swordfish size of 119 cm (applied as the equivalent dressed weight of 33 lbs or 15 kg). Consequently, smaller swordfish are mostly discarded. More than 90 per cent of the swordfish landed by United States longliners are kg T wt (~20 75 kg W wt ). The most common size of landed swordfish is 27 kg T wt (~35 kg W wt ) and, since 1992, more than 70 per cent of the landed catch has been kg T wt (~25 50 kg W wt ). Species associations Spain s longliner landings are dominated by blue shark, swordfish and shortfin mako shark. United States longliners targeting swordfish also have large incidental catches of shark, mostly blue shark and shortfin mako. Longliners and harpoon boats that target swordfish occasionally take valuable northern bluefin. Canada s and United States longline landings also include bigeye, yellowfin and albacore, which represent an increasingly important component of the catch (Dr J. Porter, 16 December 1999). At low latitudes the longline bycatch of tropical tuna, such as yellowfin, increases. We did not find references to interaction with marine mammal, such as killer whale. Swordfish fisheries 57

24 Distribution The Gulf Stream and the shelf slope or break are important features of the United States North Atlantic swordfish operations. During May June United States longliners made most sets off Cape Hatteras (34 36ºN, 72 75ºW) and south of Georges Bank. They concentrated on a narrow band along the shelf break (220 m isobath) from Cape Hatteras to Georges Bank during July August, with fewer sets near the Gulf Stream (to about 66ºW). By November and December effort was still concentrated along the shelf break, but the latitudinal distribution contracted, with most longliners fishing between Cape Hatteras and Hudson Canyon (to about 39ºN; Podesta et al. 1993). Canada s longliners currently operate from Georges Bank off Nova Scotia to the Grand Banks of Newfoundland during May November, when swordfish migrate into and adjacent to the Canada s exclusive economic zone. Fishing effort generally progresses from west to east and back again and Shark are a common bycatch of longliners targeting swordfish. from offshore to inshore along the edge of the continental shelf following swordfish movements associated with seasonal warming trends of surface water temperature and a northward movement of the edge of the Gulf Stream (Stone and Porter 1999; Beckett 1974, p. 103). In January March, swordfish are generally confined to warmer waters associated with the Gulf Stream outside Canada s exclusive economic zone and are not easily accessed by the Canada s fleet (Dr J. Porter, 16 December 1999). Originally Spain s fleet fished in the coastal waters of the southern Iberian Peninsula and the Mediterranean Sea. By 1966 Spain had established a fleet off the north-western coastline which began searching for new fishing grounds in the west. By 1985 Spain s longliners were fishing as far west as 45ºW, overlapping United States longline operations and thereby creating a continuous fishery across the North Atlantic (Miyake and Rey 1989, p. 115). By 1983 the largest of Spain s longliners were also fishing along Africa s coastline and into the South Atlantic. In 1994 and 1995 most of Spain s longline effort in the North Atlantic concentrated on two areas. The small coastal longliners from Algeciras and the larger offshore longliners from Vigo fished a large area extending off the Iberian Peninsula west to 25ºW and from 20ºN to 45ºN. Larger longliners based in the Azores fished mostly between 30 45ºN and 30 45ºW (Mejuto et al. 1997). Oceanography Along its eastern coastline, North America features extensive banks which swordfish often forage over. Waters over those banks are about 100 m deep and support substantial demersal fisheries, e.g., the Grand Banks, Scotian Shelf and Georges Banks. By comparison, the continental shelf is narrow on the other side of the North Atlantic (e.g., <50 nm along some sections of Spain s coastline). Beyond continental shelves, the North Atlantic consists of broad basins, mostly m deep. Seamounts occasionally rise from the seabed and are particularly abundant along the Mid Atlantic Ridge that bisects the North Atlantic at about 30ºW from 52ºN to the Azores (38ºN). Upper water circulation in the North Atlantic is based on a huge clockwise circulation or gyre around the Sargasso Sea. Driven by the north-east trade winds, the North Equatorial Current flows westwards along the equator. As it approaches the Caribbean it is joined by part of the South Equatorial Current and then splits into two branches. One branch flows north-westwards to become the Antilles Current off the Lesser Antilles. The other branch flows into the Gulf of Mexico. From the Gulf it returns to the Atlantic through the Straits of Florida to become the northward-flowing Florida Current. It rejoins with the Antilles Current, becoming the Gulf Stream from about Cape Hatteras (36ºN, 76ºW). It flows swiftly north-eastwards to the Grand Banks 58 Bureau of Rural Sciences

25 North Atlantic (~45ºN) and creates strong thermal gradients and induces mixing of colder, nutrient-rich waters (Pickard 1979, pp ). From Newfoundland the flow continues east and north as the North Atlantic Current. The North Atlantic Current splits, with one branch turning southwards towards Spain and North Africa to complete the North Atlantic Gyre (Pickard 1979, p. 137). Together with the system of banks off the eastern seaboard, the Gulf Stream is critical to swordfish fishing. It is known as a western boundary current, bringing a strong flow of warm water to temperate areas along the shelf break. In contrast, currents running along the western margins of continents transport cold water from low latitudes and are more diffuse. Gulf Stream water is mostly 16 20ºC, matching the preferred temperature range of adult swordfish (Pickard 1979, pp. 140, 142). The Florida Current flows over the continental shelf of the Gulf of Mexico, whereas the Gulf Stream is located outside the shelf break. The Gulf Stream is about 65 nm wide. It has a distinct boundary on its western side but, on its eastern side, the Gulf Stream is more difficult to distinguish from the Sargasso Sea. The Gulf Stream meanders and forms eddies or rings. South of the Gulf Stream those rings rotate anti-clockwise and contain cold, continental slope water. Rings north of the Gulf Stream rotate clockwise and contain warm Sargasso water. The rings are nm in diameter and as much as 3000 m deep with lifetimes of up to two years (Pickard 1979, pp. 137, 140, 143). Gear and targeting Fleets In the North Atlantic the number of United States boats permitted to fish for swordfish increased steadily, from 586 boats in 1991 to almost 1200 in 1995, primarily because discussions were under way to limit entry to this and many fisheries (NMFS 1999, p. 4-3). Permits cost nothing, had no qualifying criteria and fishers could nominate several fisheries on a single form (Dr J. Hoey, 26 February 2000). By 1997 about 1000 United States boats were permitted to fish in the Atlantic swordfish fishery (NMFS 1998, p. 246). Whereas many boats report actively fishing for swordfish, the number that actually land swordfish for more than five months each year fluctuated between 163 and 209 during the 1990s. Canada s swordfish fleet is limited to 77 longliners and not all are active in every year. Before 1993, longliners were active. As a result of the closure of groundfish fisheries nearly all licences were active from In 1998 only 49 licences were active, as a result of reduced quotas and increased opportunities to fish for other species (Porter and Allen 1998, p. 2). Longline trips are usually days in duration. In most cases boats travel to the edge of the continental shelf to fish. In 1999 there was renewed interest in fishing beyond the Grand Banks on the Flemish Cap. Canada has about 100 active harpoon licences. Spain s swordfish fishery involves three coastal fleets. One fleet is based in the port of Vigo in the north-west, another in Algeciras (south) and the third in the Canary Islands (south-west). In 1985 the Vigo fleet consisted of 125 boats, mostly GRT. The largest offshore longliners operate from Vigo, which accounts for about 80 per cent of Spain s swordfish landings. The Algeciras fleet consisted of 85 boats mostly under 50 GRT. Trips last days for the largest longliners and to 7 20 days for the smaller ones. Spain s distant-water longliners also operate in the equatorial Atlantic off Brazil and the Gulf of Guinea. In 1994 the North Atlantic fleet consisted of 110 medium and large-sized longliners that stay at sea for around 30 days and land their catch frozen. The longliners also fish for albacore tuna for 4 5 months of the year (Rey et al. 1988). In 1987 Portugal s swordfish fleet consisted of ten longliners (six about 50 GRT and the remainder GRT; Rey et al. 1988, p. 86). By 1991 the fleet had swelled to 42 longliners. Of those, four were newly built 24 m longliners based in the Azores. We did not find more up-todate information on Portugal s longline fleet. Portugal s swordfish catches in the North Atlantic declined because of quota restrictions, but the fleet is much larger and more effective than indicated by the 1991 data (Dr J. Porter, 9 November 1999). Swordfish fisheries 59

26 Fish-finding Before 1962 most Atlantic swordfish were landed by the harpoon fleet. A crew member in the crow s nest spotted swordfish at the sea surface. In the mid-1960s spotter planes came into use, allowing the fishers to cover larger areas and spot swordfish at greater depths and over a wider range of sea conditions (Brewster-Geisz 1997, p. 66). To find swordfish, longline skippers carefully monitor sea surface temperatures and water movements while searching over large areas for temperature fronts. Since the late 1980s United States and Canada s longliners have used sophisticated fish-finding technology. Loran, radar and satellite navigation are common navigational aids. Widely used since the early 1980s are onboard weather facsimile machines, access to real-time Gulf Stream and sea surface temperature data and colour-depth recorders (chromatographs) to locate the deep-scattering layer, thermoclines, seabed topography and baitfish schools (Berkeley 1989, p. 50). There is little information available on how Spain s fleet finds swordfish, but it is likely that most of the large, offshore longliners are equipped with radar and sophisticated navigational equipment and fish-finding aids. Fishing gear In the early 1960s Canada s and United States longline gear for swordfish consisted of a heavy, multistrand nylon mainline rigged shallow (Hoey and Casey 1988). The multistrand nylon branchlines were shorter than ten m, with a short monofilament leader. The gear remained virtually unchanged until the expansion of the fishery into the Florida Straits in Cuban- American fishers used traditional Cuban longlines, with very short mainlines, to successfully catch swordfish in the fast-flowing Gulf Stream and the Florida Straits (Berkeley et al. 1981). They made further gear refinements as the fishery developed (Table 4). In the early 1960s Canadian fishers used modified halibut trawl gear as pelagic longlines (Beckett 1971). Both the mainline and branchlines were made of tarred, multistrand nylon and the branchlines were spliced directly to the mainline. They first used monofilament nylon in the late 1960s when they attached short (~38 cm) leaders to the multistrand branchline above the hook. In the late 1970s following the eight-year closure of the fishery due to mercury restrictions, Canada s longliners used branchlines consisting of an upper tarred multistrand section, with a clip for attaching it to the mainline and a lower monofilament nylon leader. The monofilament leader made up one-half (~1.8 m) of the overall length of the branchline (~3.6 m) and was considered to yield better catch rates because it was less visible to swordfish and other pelagic species. In the 1980s longliners began to use monofilament nylon for both sections of the branchline and for the mainline on the assumption that this would further improve catch rates (Dr J. Porter, 16 December 1999). The introduction of lightsticks was a significant innovation for targeting swordfish. Professional charter boat operators and recreational anglers in Florida were the first to experiment with lightsticks to catch swordfish during Their success prompted several longliners to try them and soon they were in widespread use (Berkeley et al. 1981). Other important developments were the use of long, single strand monofilament leaders, combined with longer branch and buoy lines that allowed the gear to be fished deeper. The mainline was also shorter, with fewer hooks, spaced wider apart. Those innovations doubled swordfish catch rates in the Straits of Florida (Hoey et al. 1989). By 1980 monofilament mainlines were popular in southern areas and were widespread by the mid-1980s (Berkeley et al. 1981; Dr J. Hoey, 1 March 2000). Spain s longlines are of similar lengths to those used by United States longliners, but the Spanish set hooks closer together and very shallow (Folsom 1997, p. 102). During Japan s longliners made significant changes to their longline gear. About per cent of the fleet introduced monofilament mainlines, branchlines and leaders, or synthetic, thin braided nylon lines. Only 34 per cent of the fleet used nylon mainlines in 1994 increasing to 76 per cent by 1996; 41 per cent of the fleet used nylon branchlines in 1994, increasing to 77 per cent in 1996 (FAJ and NIFSF 1998, pp ). But there are reports that many longliners reverted back to traditional gear after Bureau of Rural Sciences

27 North Atlantic Fishing practices Fishing practices are now much the same for most of the North Atlantic fleets targeting swordfish. Because of the large number of hooks and the lengthy soak and haul times, longliners make only one set each day. Usually longliners set their gear at dusk and commence hauling before dawn. The early Cuban longliners used an interesting technique that, while only applicable to gear with short mainlines, produced a twofold benefit that is relevant to the modern swordfish fishery. The fishers tended their lines during the night, using kerosene lamps mounted on floating platforms. When a fish was hooked, the increased pressure on the line flipped the light platform, extinguishing the light. This alerted crew members, so that each fish could be hauled aboard as it was caught and the hook redeployed. Catch rates were high because the line was intensively tended and the light platforms also attracted swordfish. Swordfish were freshly caught and undersized swordfish could be released while they were still alive (Berkeley et al. 1981, p. 17). Discarding of target species ICCAT s minimum size limit (see p. 74) contributed to increases in swordfish discards from 6 per cent in 1990 to more than 35 per cent in 1993 (Cramer et al. 1995, p. 149). A 1995 study of bycatch in western Atlantic pelagic longline fisheries found that dead discards comprised about per cent (by number) of the total catch (Hoey 1995). Discards are predominantly small swordfish, however, so they account for only about 15 per cent of the total weight of the catch. United States logbook reports of swordfish discards have not been particularly accurate. Several analyses of observer data have shown that observed catch rates of discarded swordfish were higher than logbook reports (Cramer et al. 1995, p. 149). Only the United States ( ) and Canada ( ) report estimates of dead discards, which are based on observer data. In 1997 logbook and observer data revealed that United States longliners discarded about dead swordfish. Other species discarded included an estimated 9284 pelagic shark, 5871 coastal shark, 3658 white marlin, 2739 sailfish and 2190 blue marlin (Cramer and Adams 1999, p. 10). Spain s longliners operating east of 20ºW land amounts of shark (~70% of the total catch, by number). The shark are mostly blue shark and shortfin mako (Buencuerpo et al. 1998, p. 669). Shark, including blue shark, are regularly landed and marketed in southern Spain. They attract low market prices, however, so many are discarded, particularly by larger longliners when storage space is limited. Discarding of swordfish is uncommon on Spain s longliners. About two per cent of the catch are eaten on board or are shark-damaged and discarded and not counted in the catch statistics (ICCAT 1995). Swordfish fisheries 61

28 Table 4. Summary of boat types and longline gear used in the North Atlantic swordfish fishery. Characteristic Cuban-Americani Traditional New Englandii Floridaiii Early Spanish surface longlineiv Caribbeanv Asian distant-watervi Average boat size 8 12 m LOA m LOA 8 20 m LOA GRT (up to 750 GRT) in the north; ~50 GRT in the south LOA m LOA; GRT Average trip length one night at least 2 weeks one or two nights 1 4 weeks, depending on boat size up to 18 months Type of gear two strands of kg braided monofilament together, deployed from baskets monofilament mainline stored on hydraulic reels braided nylon or lead-core polypropylene mainline, some use hydraulic reels, some baskets braided polyethylene mainline 320 kg monofilament mainline stored on hydraulic reels many now have monofilament mainlines stored on hydraulic reels Length of mainline 3 7 km km km km km km Diameter of mainline 5 8 mm Distance between floats <100 m ~270 m ~120 m Length of floatline ~12 m ~12 m ~15 m Length and type of branchline 2 5 m two-ply monofilament braided to the mainline <10 m monofilament nylon branchline and leader attached to mainline with metal clip m snap-on braided branchline with short monofilament leader multifilament nylon braided to the mainline m depending on water temperature Hook type 3/0 or 3-1/2/0 shark hooks no. 8 or 9 mustad J-hooks, slightly offset 9/0 12/0 big game hooks 8/0 12/0 big game hooks Hook spacing m 30 m (~35 hooks per km)50 85 m (~12 18 hooks per km) ~30 m m (~6 8 per km) 62 Bureau of Rural Sciences

29 North Atlantic Table 4. Summary of boat types and longline gear used in the North Atlantic swordfish fishery (cont d) Number of hooks per set average but occasionally set up to average (range depending on boat size) >500 ~2000 Type of lightsticks not used various brands Cyalume snap type chemical sticks on every branchline, or every second branchline not usedcyalume snap type chemical sticks attached 3 5 m above the bait Type of bait silver mullet with Spanish sardine or thread herring squid or Atlantic mackerel squid or Atlantic mackerel mackerel or sardine squid or Atlantic mackerel Swordfish fisheries 63

30 Assessment Data collection The United States s longline and harpoon boats have provided information on catches to the United States National Marine Fisheries Service (NMFS) in logbooks that have been mandatory since 1986 (Hoey et al. 1995, p. 248). Since 1998 logbooks have provided information on boat characteristics and, for each longline set, location, set and haul times, sea surface temperature, gear configuration, the number of each species retained and the number discarded (Cramer and Adams 1999, p ). In 1999 NMFS required logbooks to be completed within 48 hours of the day s fishing activities and lodged before offloading of the boat (NMFS 1999, p ). Swordfish dealers must hold a permit and submit fortnightly reports to NMFS detailing the weights of all swordfish purchased and registration details of each corresponding boat. Those details are used to cross-check against logbooks (Cramer and Adams 1999, p. 8). NMFS began placing contracted and NMFS observers on United States longliners in The selection of boats is random, with coverage set at 5 per cent of the number of sets reported by the longliners in the previous season. They observed 2857 sets in the United States Atlantic fishery during (NMFS 1998, p. 252). In 1999 NMFS made vessel monitoring systems (VMS) 3 mandatory for all United States-flagged longliners operating in the North Atlantic (NMFS 1999, p ). VMS must be used to submit hourly position reports to NMFS. VMS thereby facilitates the enforcement of the time and area closures that were also implemented in NMFS is developing a system for reporting of catch and effort data through VMS (NMFS 1998, p. 252). The VMS requirements are due to become effective in mid-2000 (Dr J. Hoey, 26 February 2000). Canada s longliners have voluntarily completed logbooks since the fishery began in the early 1960s. In 1994 the Department of Fisheries and Oceans (DFO) introduced a dockside monitoring program (DMP) for the swordfish and northern bluefin fleets (Porter and Allen 1998). The Program requires each boat to provide logbook data to an independent, industry-funded monitoring company immediately upon entering port. Program staff weigh each fish as it is landed. Boats cannot leave port for their next fishing trip until data requirements are fulfilled. The dockside monitoring program has ensured 100 per cent compliance, which was mostly less than 50 per cent before the implementation of the Program. Canada established a VMS system in 1999 to meet an ICCAT recommendation (Dr J. Porter, 16 December 1999). The Instituto Español de Oceanografía has collected data on Spain s fishing effort and the number and weight of swordfish caught since Institute staff collect size data through port sampling, voluntary logbooks and observer placement. They monitor landings for virtually every day during the fishing season. When a trip is selected, 100 per cent of the catch is sampled. They measured about 50 per cent of the swordfish landed by Spain s boats in Discarding of small swordfish at sea is not considered a problem because Spain s market prefers small swordfish. There are also more than observer records of swordfish size and sex (Mejuto et al. 1988; Dr J. Mejuto, 26 April 1999). Research ICCAT is a multinational body relying on member nations to collect data and to undertake research on the stocks. North Atlantic swordfish are the most extensively studied swordfish stock. Our review of swordfish biology (p. 5) presents results of many studies of North Atlantic swordfish. The present section highlights some of the current areas of research on North Atlantic swordfish. In 1999 SCRS proposed that it use the next inter-sessional meeting to improve basic biological knowledge of swordfish. SCRS s extensive use of virtual population analysis has necessitated studies of swordfish age and growth. SCRS has a long history of developing abundance indices from commercial fishing catch and effort data. It has now incorporated sex ratio information and 3 Inmarsat C units that submit real-time data direct to NMFS using on-board satellite transceivers. 64 Bureau of Rural Sciences

31 North Atlantic some environmental factors in several models. SCRS is also keen to obtain more information on swordfish reproduction, particularly the seasonal and geographical distribution of spawning grounds and swordfish fecundity (ICCAT 1999c, p. 24). Scientists representing contracting parties of ICCAT have investigated the stock structure of swordfish in the Atlantic Ocean and Mediterranean Sea and compared Atlantic swordfish with other areas, such as the Indian Ocean. They have reported significant differences in allele frequency between swordfish sampled in the Mediterranean, north-western Atlantic, southern Atlantic and Indo-Pacific. Similarly, differences in frequencies of other alleles were also observed between the north-western Atlantic Mediterranean and the tropical Atlantic and the temperate southern Atlantic (ICCAT 1999c, p. 4). The 1999 SCRS meeting reviewed results of two tag recapture studies. During Canada tagged and released 357 juvenile swordfish (<119 cm) in Canada s exclusive economic zone. Of the seven recaptures, fishers reported two from the central northern Atlantic, four within Canada s zone and one east of the Bahamas (ICCAT 1999c, p. 3). During Spain s longliners recaptured 47 swordfish that had been tagged and released by various agencies. Most recaptures were reported from the North Atlantic between 20 45ºN and 5 60ºW. Those recaptures were from swordfish released north of 10ºN. There were only three returns from the southern Atlantic, which were swordfish that had been tagged and released south of 20ºS (ICCAT 1999c, pp. 3 4). Stock assessment Fishery indicators In early assessments of the North Atlantic swordfish stock, SCRS cited reductions in average swordfish size as evidence that the stocks were in decline. For instance, size frequency data from Spain s and United States longliners showed no clear trend during , but then size had declined by Hoey and Mejuto (1991) suggest that the decline was possibly a response of the population to increasing fishing pressure on all ages of swordfish. In a 1994 assessment, SCRS found that many factors influenced the average size of swordfish in the North Atlantic (ICCAT 1995). The almost continual expansion of the fleets, the effects of the minimum size rule on landings and size samples (as the majority is derived from port sampling) and the 1970s mercury restrictions are likely to have had an effect. Market forces, at least in the United States and Canada, also exert a strong influence on the size composition of swordfish catches (Dr J. Porter, 9 November 1999). Therefore, historical changes in the size distribution of swordfish caught in the North Atlantic may not be an accurate indicator of the effects of fishing pressure on the stocks. For all nations fishing in the Atlantic, about 19 per cent (by number) of landed swordfish in 1998 were less than ICCAT s minimum size (125 cm). 4 Adding estimates of discards raises the percentage of swordfish less than 125 cm to about 23 per cent. The size data provided by Canada, Spain and the United States to ICCAT, suggest that longliners caught large numbers of small swordfish in Spain s catches of small (<125 cm) swordfish increased to 37 per cent in 1998, perhaps reflecting increased recruitment rather than changes in fishing patterns. Canada (21%) and the United States (32%) estimated large numbers of small swordfish in their 1998 catches (ICCAT 1999b, p. 4). In Spain s North Atlantic fishery, nominal catch rates show no clear trend during the 1970s and 1980s, but those levels were only maintained by the movement of the fleet to new fishing grounds where higher catch rates masked the declining trend in catches in the earlier fishing areas (Berkeley 1989, p. 50). Similarly, rapid improvements in fishing gear and efficiency in the northwestern Atlantic have made it difficult to quantify real catch rate trends. As such, nominal catch rates have not been accurate indicators of relative abundance. 4 ICCAT provides two options for applying minimum size limits: 125 cm with a 15 per cent tolerance or 119 cm with zero tolerance. Only Canada and the United States have applied the zero tolerance option. Swordfish fisheries 65

32 In 1999 SCRS incorporated standardised catch rates for ages 1 5+ for the unisex virtual population analysis (aged using the Gompertz growth equation) and 1 9+ for the base case sexspecific virtual population analysis (aged using the Ehrhardt growth equations; Ehrhardt et al. 1996). It applied that procedure to data sets for each of the major fleets: Japan (ages 3 9+, ), Spain (ages 1 9+, ), United States (ages 1 9+, , excluding ages 1 and 2 from ) and Canada (ages 2 9+, ). SCRS used two methods to split the catches into age classes: applying the Gompertz growth curve for both sexes combined; and applying Ehrhardt s sex-specific growth curve (ICCAT 1999c, pp. 7 8). For most fleets the catch rates of swordfish younger than age 4 show no clear trend. All fleets show some year effects not explained by the models and this issue has become a priority for the SCRS to resolve. Catch rates for ages 4 and the 5+ age class generally declined over most of the series for all four fleets. In 1999 catch rates showed signs of stabilising or improving on levels reported in previous years (ICCAT 1999b, p. 2). In particular, standardised catch rates of age one swordfish for Spain s fleet increased in 1997 and Stock assessment models Since 1986 SCRS has assessed the status of North Atlantic swordfish nine times and reviewed catch levels and catch rates in between assessment years (Table 5). Initially, when the stock was being fished down, SCRS did not have a reliable estimate of the maximum sustainable yield (Dr J. Hoey, 26 February 2000). The progressive introduction of a wide range of models, refinements to models and improvements to data have made assessments more sophisticated and reliable and allowed the review of past assessments. Since 1988 SCRS has used age-structured virtual population analysis, introducing surplus production models (in 1990), non-equilibrium production models (1991), age-structured production models (1996) and sex-structured and agestructured virtual population analyses (1996; Restrepo 1998, p. 518). In 1998 work focused on methods to incorporate sex-ratio at size information into the agestructured sequential population analysis. The geographical areas used in analysis were characterised as representing sex-ratio at size profiles indicative of spawning, feeding and transitional aggregations of swordfish (Dr J. Porter, 21 June 2000). The virtual population analyses project stock dynamics 15 years into the future. SCRS has explored the sensitivity of assessments to various assumptions and uncertainties in the data, e.g., catch rate series, recruitment and estimates of discards. It has used Bayesian approaches to summarise the probability distributions for those uncertainties (ICCAT 1999c, pp. 12, 17). SCRS also developed a standardised biomass index from catch rate information derived from data sets provided by United States, Canada, Japan and Spain. They standardise nominal catch rates using generalised linear model protocol including year, area, quarter, a targeting variable and a nation-operation variable accounting for gear or operational differences thought to influence swordfish catchability. The biomass index has generally declined over the time series ( ), but there has been a slight increase in the most recent years. The production model is used mostly for management advice because it gives a direct estimate of maximum sustainable yield, which is ICCAT s management objective (Dr J. Porter, 16 December 1999). The earliest assessments suggested that fishing mortality rates had increased in the late 1980s, soon after the peak in catch levels. The stock was at about 85 per cent of carrying capacity when the longline fishery started in the early 1960s. Longlining fished the stock down, then it rebounded slightly during the mercury closure during the 1970s and then continued to decline after the fishery reopened. SCRS s 1990 assessment showed that the stock fell below the optimum biomass in about 1989 (Dr J. Hoey, 26 February 2000). The 1992 assessment predicted a slightly improved stock status, partly as a result of underreporting of catches by several nations. Subsequent assessments, using corrected data, highlighted a worsening status for North Atlantic swordfish (Dr J. Hoey, 26 February 2000). In particular, the 1996 assessment indicated that the stock had continued to decline despite reduced landings since the 1987 peak catch. Although some nations reduced their catch levels by substantial amounts, this did not reduce the overall fishing mortality rate because recent landings continued to exceed 66 Bureau of Rural Sciences

33 North Atlantic surplus production. Declining catch rates in several fleets reflected the decline in stock size (ICCAT 1997a, p. 170). Status The stock assessment models applied to North Atlantic swordfish by SCRS in 1999 all show that the stock is below the level that would provide maximum sustainable yield ( B MSY ; Table 6). But, compared with the 1994 and 1996 stock assessment, the 1999 assessment provides a slightly more optimistic outlook for North Atlantic swordfish. SCRS attributed the better outlook to improvements in the data quality and analytical treatment. Retrospective analyses revealed that, had the revisions and data treatment used in 1999 been available in 1996, the assessment would have been much the same (ICCAT 1999c, p. 12). The 1999 virtual population analysis extended age classes, so that the plus groups consisted of age 9+ female swordfish and age 5+ males (ICCAT 1997c, p. 14). The 1999 assessment suggests that the decline in swordfish biomass has slowed (Figure 14). The sequential population analysis shows that recruitment of age one swordfish increased in the early 1980s before declining during The analysis then shows strong recruitments of age one swordfish in 1997 and 1998 (ICCAT 1999b, pp. 2 3). The trends for ages 2, 3 and 4 are similar to the trend for age 1 swordfish with the appropriate time lags, but the pattern is less pronounced. However, the estimated abundance of age 5+ swordfish at the beginning of 1999 was about one-third of the 1978 level (ICCAT 1999b, p. 2). The estimated fishing mortality rate increased during for most ages of swordfish (Figure 13). The 1998 fishing mortality rate was 1.34 times the fishing mortality rate at maximum sustainable yield. SCRS predicted that, if the trends in fishing mortality were to continue, the biomass of adult female swordfish would be reduced to about 8 per cent of the level at maximum sustainable yield. This is well below the level that is considered to result in risks of recruitment over-fishing in other fish species (ICCAT 1999b, p. 2). Swordfish fisheries 67

34 Table 5. Summary of SCRS ( ) and Southeast Fisheries Center (1986) stock assessments of North Atlantic swordfish. Except for relative biomass, all values are metric tonnes (t) per year. Year Features of assessment Major conclusion(s) MSY 80% confidence interval for MSY Relative biomass (Byr/BMSY) 15-year replacement yield Subsequent annual catch 1986i Virtual population analysis tuned to catch per trip of swordfish greater than 130 lbs catches exceeded stock production. Spawning stock biomass is low Virtual population analysis tuned with indices of abundance derived from Spanish and United States data. Modest to very large increases in fishing mortality on all ages during % decline in spawning biomass. Recruitment increased substantially during Introduced surplus production modelii. Continuing decline in the abundance of adults (age 5+) during Fishing mortality needs to be halved for stock to recover to the biomass corresponding to MSY t 1992iii Introduced non-equilibrium surplus production model. Catch data extended to Population decline slowed or stabilised. Recruitment increased in and declined in Fishing mortality rates declined from the peak levels of t iv4 277 t 84% t t 1994 Fishing mortality for all ages increased throughout the 1980s, then stabilised before increasing in fishing mortality rate almost twice the rate predicted to produce that yield. Catch reductions not reducing fishing mortality t t 68% t t Bureau of Rural Sciences

35 North Atlantic Table 5. Summary of SCRS ( ) and Southeast Fisheries Center (1986) stock assessments of North Atlantic swordfish. Except for relative biomass, all values are metric tonnes (t) per year (cont d) 1996 (cont) Catch data extended to Preliminary sex-specific VPA. Abundance of age 5+ swordfish halved during Recruitment of age 1 swordfish increased in early 1980s, then remained stable to 1989 before declining slightly to Fishing mortality on ages above 1 reached or exceeded the peak levels during Biomass would not reach the level corresponding to maximum sustainable yield without substantial reductions to catches. Biomass of adult females will be reduced to 2% of the maximum at equilibrium t t 58% t t 1999 Improved data quality and analytical treatment. Extended age classes ( plus group consisted of age 9+ females and age 5+ males). Introduced sex specific sequential population analysis (SPA). Decline in biomass slowed or arrested. Strong recruitments of age 1 swordfish in 1997 and Abundance of age 5+ swordfish one-third of the 1978 level. Fishing mortality rate increased during for most ages of swordfish. Biomass of adult females will be reduced to 8% of the maximum at equilibrium t t 65% t Swordfish fisheries 69

36 SCRS s 1999 estimate of maximum sustainable yield ( t) for North Atlantic swordfish was similar to previous assessments (Table 5). Since 1983 catches have been less than t in only three years (1984, 1997 and 1998). Although the decline in swordfish biomass has been slowed or arrested, the biomass is estimated to be only 65 per cent of the biomass needed to produce maximum sustainable yield. An annual catch of not more than t (the previously existing total allowable catch) was required to rebuild the stock to that biomass within 15 years. Catches above the replacement yield ( t) are unlikely to allow the stock to rebuild. The preliminary estimate of the 1998 catch ( t) was above that level, but catches in 1999 are expected to be lower than in 1998 because of current management regulations (ICCAT 1999b, p. 2). SCRS recommended that, for the stock to rebuild to levels that would support maximum sustainable yield within ten years with a 50 per cent probability, fishing nations limit total swordfish catches to t per year (Figure 14). At the 1999 catch limit of t, there would be a greater than 50 per cent chance of reaching those levels in 15 years. But, the stock would not reach those levels if catches continued at t per year (the 1999 catch limit plus ten per cent for over-catch). More effectively implementing restrictions on catching small swordfish would provide further gains in stock rebuilding. Specifically, the recruitment of age 1 swordfish in 1997 and 1998 would contribute to increases in spawning biomass if those year classes are not heavily harvested before they reach maturity (ICCAT 1999b, p. 4). SCRS has provided clear advice that the North Atlantic swordfish stock is below the level that would produce maximum sustainable yield. Catch levels must not exceed t if the stock is to have an acceptable probability of rebuilding in a reasonable time frame. This highlights the problem of a total allowable catch: a constant total allowable catch applied to a continually decreasing stock will result in everincreasing levels of fishing mortality and, therefore, overexploitation. SCRS suggested that target fishing mortality rates, rather than total allowable catches, were less risky for rebuilding the swordfish stock. Target fishing mortality rates can be translated into total allowable catches and adjusted based on the status of the stock. Because the total allowable catch corresponding to a particular level of fishing mortality is dependent on the current stock size, management measures are likely to need adjustments after each assessment to maintain the target fishing mortality rate. Standardising catch rates Catch rates have been used to monitor fluctuations in swordfish abundance. Nominal or raw catch rates need to be standardised for changes caused by various factors (e.g., fishing gear, targeting, area and time of year) to develop indices that represent the abundance of the stock. When longliners use lightsticks, for example, nominal catch rates of swordfish might be double those for tuna sets. The catch rates would need to be adjusted for the effects of lightsticks if they are to reflect the abundance of swordfish in the area. The assumption that catch rates represent relative abundance is fundamental to production models and virtual population analyses. The reliability of stock assessments are dependent on that assumption being true. Nominal catch rates are standardised using generalised linear models (GLMs) or generalised additive models (GAMs). SCRS has highlighted the pitfalls of using constant total allowable catch to manage North Atlantic swordfish. An alternative management strategy, fishing mortality-based ( F-based ) management, is difficult to implement especially in an international fishery. More recent SCRS advice has concentrated on reducing catches, controlling catches above the total allowable catch and protecting small swordfish (Dr J. Porter, 16 December 1999). Reliability of the assessment There is a considerable variance in the form and reliability of catch and size data submitted by members to ICCAT for assessments. Major nations are submitting data on a finer scale and ICCAT now has the capability to partition the catch by five-degree squares. ICCAT s catch data are limited to large-scale areas, which hampers the interpretation of fine-scale differences in catch rates. SCRS recommended that members report data by five-degree square, which is still a fairly imprecise unit of resolution. Recent changes in targeting create further uncertainties in the assessment (ICCAT 1999c, p. 19). Targeting data were presented at the 1999 assessment and 70 Bureau of Rural Sciences

37 North Atlantic SCRS needs to develop a process for dealing with targeting when standardising catch rates (Dr J. Porter, 16 December 1999). SCRS suggests that ICCAT landings data should be considered minimum estimates because of unreported landings by non-members and from boats flying flags of convenience and through bycatch in other fisheries (ICCAT 1999b, p. 1). Assessments also require improved estimates of the number of undersized swordfish caught and discarded and likely to die (ICCAT 1999c, p. 23). Canada and the United States estimate length from dressed weight at landing, but those conversions may affect the total variance of length-atage data. The Spanish provide data for swordfish landed whole, but there are suggestions that unreported catches are landed at various ports. Japan s data are from onboard industry sampling (Porter 1997). Provision of size data by other nations (e.g., Portugal and Taiwan) improved during the late 1990s, reducing the need for substitution of size data between fleets. Table 6. Summary of the SCRS 1999 assessment of North Atlantic swordfish (ICCAT 1999b, p. 5). Reference point Value Range Maximum sustainable yield a t t c Preliminary 1998 yield t Replacement yield (2000) t t Relative biomass (B 1999 /B MSY ) Relative fishing mortality: F 1999 /F MSY a F 1999 /F max b F 1999 /F 0.1 b Key B: biomass MSY: maximum sustainable yield F: fishing mortality F max : the fishing mortality that maximises yield per recruit at equilibrium F 0.1 : the fishing mortality rate at which the increase in equilibrium yield per recruit in weight for an increase in a unit of effort is ten per cent of the yield per recruit produced by the first unit of effort on the unexploited stock a Production model based on data. b Sequential population analysis based on data for female swordfish only. c 80 per cent confidence interval. Swordfish fisheries 71

38 MSY Year 0.0 Figure 13. Annual estimates of relative fishing mortality (the ratio of current fishing mortality to that at maximum sustainable yield) for North Atlantic swordfish, estimated by the ASPIC production model (grey lines are Bayesian 80 per cent confidence intervals for the estimates; ICCAT 1999c). Also shown is the fishing mortality at maximum sustainable yield (MSY; broken line) MSY Relative biomass (B/Bmsy) Year 0.0 Figure 14. Annual estimates of relative biomass (the ratio of current biomass to that at maximum sustainable yield) for North Atlantic swordfish, estimated by the ASPIC production model (grey lines are Bayesian 80% confidence intervals for the estimates; ICCAT 1999c). Also shown is the biomass at maximum sustainable yield (MSY; broken line) and relative biomass projected to the year 2010 based on constant scenarios of annual catch (0 t, t, t, t and t). 72 Bureau of Rural Sciences

39 North Atlantic Another area of uncertainty is with the biology and behaviour of swordfish, particularly validated growth curves (Porter 1997, pp ). SCRS is also concerned about the estimate of length at infinity ( L and the growth coefficient, K ) for male swordfish from the growth curve used to convert lengths to age (ICCAT 1999c, p. 24). Stock structure remains a concern to stock assessment and management of Atlantic swordfish. Genetic analyses (p. 65) indicate a distinct Mediterranean stock and separate North Atlantic and South Atlantic stocks. However, the precise boundaries between stocks are uncertain (ICCAT 1999b, p. 1).and we note that swordfish catches for 1996 show no obvious discontinuity across equatorial waters of the Atlantic. Consequently, assessments and fishery management need to take into account the full range of uncertainty regarding the boundaries of those stocks. Preliminary studies had shown that there may be a causal link between the North Atlantic Oscillation (NAO, p. 88) index and North Atlantic swordfish recruitment. The relatively low levels of the index during corresponded to a 50 per cent increase in catch rate of age one swordfish (ICCAT in press a, p. 6).However, further work is required on the relationship between recruitment and environmental conditions because the correlation was not consistent, for example, between fleets and between different forms of the index (ICCAT 1999c, p. 21). Management Institutions Comprised of 27 contracting parties, ICCAT coordinates the assessment and management of Atlantic and Mediterranean tuna and tuna-like species, including swordfish (ICCAT 1999a). It recommends regulations but has no regulatory authority. Contracting parties may adopt those recommendations and implement them within their domestic management arrangements. For assessment and management of swordfish, ICCAT treats the north and South Atlantic as separate management units, with 5ºN latitude as the boundary between the two units. Genetics studies have shown that Mediterranean swordfish are distinct from Atlantic swordfish. Mediterranean swordfish are assessed and managed separately from the Atlantic, although swordfish are known to move at least within the Strait of Gibraltar and the status of those swordfish is unclear. The National Marine Fisheries Service (NMFS) is a United States government agency responsible for ensuring that United States s boats fishing for swordfish in the North Atlantic comply with ICCAT management recommendations. ICCAT regulations are implemented in the United States under the Atlantic Tuna Convention Act (1974) and the Magnuson Stevens Fishery Conservation and Management Act (1996). The Magnuson Stevens Act requires fishery managers to halt overfishing, rebuild overfished stocks, minimise bycatch and bycatch mortality and protect essential fish habitat. In Canada, the Department of Fisheries and Oceans sets multiyear management plans for swordfish, based on the ICCAT recommendations and in conjunction with fishers and buyers through the Atlantic Large Pelagics Advisory Committee. In Spain, the Administrative Service of Fisheries is responsible for managing the longline fishery. Its links with the Instituto Español de Oceanografía, which monitors the fishery, are weak (Dr J. Mejuto, 26 April 1999). Under the mandate of the Common Fisheries Policy the European Union has negotiated fishery agreements on behalf of member states since Although not yet a formal member of ICCAT, the European Union has participated in ICCAT meetings since Swordfish fisheries 73

40 Objectives ICCAT formulates management recommendations for the North Atlantic swordfish stock with the primary objective of achieving maximum sustainable yield 5 (ICCAT 1999a). It is noteworthy, however, that the Food and Agriculture Organization has stressed that maximum sustainable yield should not be considered a conservative biological reference point. Rather, it represents a limit reference point. The use of more conservative reference points should provide a greater cushion against overfishing. ICCAT recommends total allowable catches with the objective of a 50 per cent chance of rebuilding the North Atlantic swordfish stock to the level at maximum sustainable yield ( B MSY ) in ten years (ICCAT 1999c, p. 17). Listing as an endangered species The World Conservation Union regularly assesses the status of species in the wild and maintains lists of the world s threatened species (IUCN Red Lists). National governments, regional organisations and international organisations use Red Lists as a guide to the conservation status of species. The lists also identify problems where remedial action may be needed. In 1996 the World Conservation Union added the North Atlantic swordfish stock to its Red List of Threatened Animals under the category of Endangered (A1b, d). The Endangered category is for species facing a very high risk of extinction in the wild in the near future. Worldwide, it classifies swordfish as Data Deficient ; there is inadequate information to assess its risk of extinction based on its distribution or population status. Under the Magnuson Stevens Fishery Conservation and Management Act, the 1999 Final Fishery Management Plan for Atlantic Tunas, Swordfish and Sharks (the FMP ) has the objective of managing the fisheries for continuing optimum yield to provide the greatest overall benefit in terms of food production and recreational opportunities, while preserving traditional fisheries and protecting marine ecosystems (NMFS 1999, pp to 1-14).The plan defines optimum yield as the biological maximum sustainable yield reduced by any relevant social, economic or ecological factors. Canada s swordfish fishery is managed under the Fisheries Act, Canadian Atlantic Swordfish (Xiphias gladius) Integrated Management Plan. The Department of Fisheries and Oceans updates the plan in conjunction with the Atlantic Large Pelagics Advisory Committee. The objectives of the plan are to minimise the mortality of small swordfish and reduce the bycatch of northern bluefin tuna and shark (DFO 1997). Management measures In 1991 SCRS declared that the current catch levels of swordfish could not be sustained in the long-term (ICCAT 1999a). ICCAT immediately recommended that fishing mortality of swordfish larger than 25 kg be reduced by at least 15 per cent, by limiting the catch to below 1988 levels for major fishing nations. They also recommended minimum size regulations, prohibiting the landing of any North Atlantic swordfish under 25 kg (125 cm LJFL, with 15% allowable tolerance). Some scientists consider that stock status would have been worse if nations had not implemented the minimum size regulation (Dr J. Porter, 16 December 1999), whereas others doubt the effectiveness of the regulation. United States longliners showed the only marked decrease in the landings of small swordfish, which was offset by increased discards of dead swordfish (Cramer 1996a). In 1990 the greatest number of swordfish landed by United States' boats was in the kg class. Following the implementation of the minimum size rule, the most frequently landed size shifted to kg. The mortality of juvenile swordfish might not have been reduced, however, because longline-caught swordfish have a low survival rate (Cramer et al. 1995). The size regulation did not cause longliners to move away from areas where undersized swordfish congregated. Only in the southern Caribbean and the Venezuela Basin a known swordfish nursery area did effort decline between 1990 and Closer to the United States coastline, where catch rates of juvenile swordfish tend to be high, there was no change in the pattern of fishing whatsoever (Cramer 1996a). 5 The ICCAT convention refers to maximum sustainable catch instead of maximum sustainable yield. 74 Bureau of Rural Sciences

41 North Atlantic SCRS has applied sensitivity analyses to determine the effectiveness of minimum size regulations in terms of the gains in yield per recruit (YPR) and spawners-per-recruit (SPR; Mace 1995). Unless small swordfish can be avoided rather than discarded, or survival rates of discarded swordfish improved, SCRS showed negligible benefits for a minimum size that is well below the size at maturity. In 1999 SCRS again expressed concern over the continued high catches of small swordfish and emphasised that increases in yields could accrue only if the catches and discards of small swordfish were further reduced (ICCAT 1999b, p. 4). SCRS has repeated its concern over small swordfish at every meeting since 1991, yet fishing nations have largely ignored that advice. During the 1990s SCRS consistently highlighted the need for substantial reductions in catch levels. In 1994 ICCAT set North Atlantic catch quotas for 1995 and 1996 for Canada, Spain, Portugal and the United States. All other nations were asked to limit their North Atlantic catches during to below 1993 or 1994 levels (whichever was higher). In 1996 ICCAT recommended that the total catch of North Atlantic swordfish by member nations be reduced to t for 1997, t for 1998 and t for They introduced specific quotas for those nations that have traditionally harvested the largest North Atlantic catches (the United States, Canada, Spain, Portugal, Japan and Bermuda). ICCAT requested all other nations fishing in the North Atlantic keep their catches below the 1996 levels (as reported to ICCAT in 1997). It amended the 1991 minimum size regulation to 119cm LJFL with no tolerance or 125 cm with 15 per cent tolerance. ICCAT set the total allowable catch for North Atlantic swordfish at t for Reported landings ( t) exceeded the total allowable catch by 6 per cent. When estimated discards are added to landings, the catch ( t) exceeded the total allowable catch by 11 per cent. Of the six country-specific quotas, Bermuda, Canada, Portugal, Spain and the United States were within one per cent of their 1998 allocation. Catches of other fishing nations exceeded the catch limit recommended by ICCAT (ICCAT 1999b, p. 3; ICCAT 1999d, p. 9). Until the late 1990s many member nations had not observed ICCAT regulations for swordfish and the regulations had been ineffective at conserving Atlantic swordfish stocks (Porter 1997). In 1997 the Compliance Committee of ICCAT introduced penalties (e.g., quota penalties and trade restrictions) for both member and non-member nations that do not follow ICCAT regulations in an attempt to improve compliance (ICCAT 1999a). United States The United States established its Fishery Management Plan for Atlantic Tunas, Swordfish and Sharks in April 1999 (NMFS 1999). The Plan incorporates all previous ICCAT regulations for swordfish and introduces several new regulations to halt overfishing. A fundamental objective of the fishery management plan is to rebuild the Atlantic swordfish stocks. To meet this objective, managers must determine when stocks are overfished and in need of rebuilding to the level that can support maximum sustainable yield. Fishery managers must identify reference levels to indicate when fishing mortality is greater than fishing mortality at maximum sustainable yield ( F MSY ) and for when the Swordfish fisheries 75

42 stock biomass is less than that at maximum sustainable yield ( B MSY ). They are bound to take management action when those criteria indicate that the stock is overfished (as is currently the case for North Atlantic swordfish). The United States NMFS officially closed the Atlantic swordfish fishery from 12 April to 1 June to meet its ICCAT quota in The 1999 plan also introduced time and area closures to protect aggregating juvenile swordfish for the first time. In 2000 NMFS established several large closed areas to protect juvenile swordfish in the Florida Straits, north-eastern Gulf of Mexico and from the Texas Mexico border to the Florida panhandle. The package also initiated a longliner buyback program and bycatch and mortality reduction research (Anonymous 2000, p. 13A). In 1995 NMFS had proposed to limit access to the swordfish fishery in an attempt to reduce the large number of latent permits in the fishery. The 1999 fishery management plan formally established a two-tiered, limited access system, consisting of directed and incidental permits, based on catch history. NMFS will issue a swordfish permit to boats that can prove a catch history. Boats that can prove incidental bycatch will be issued a permit to land a restricted number of swordfish each season. Under the new regulation, only 198 boats are eligible to hold a swordfish permit after 1999 and a further 218 boats qualify for an incidental permit. In 1999 the United States Department of Commerce introduced a regulation banning imports of Atlantic swordfish weighing below 33 lbs T wt (15 kg T wt or ~20 kg W wt ) to facilitate the implementation of its minimum size rule. As part of the ban, imports of swordfish require a Certificate of Eligibility verifying origin and size. Food and Drug Administration regulations for mercury in swordfish require that nations exporting swordfish to United States markets reduce the number of very large swordfish in their shipments. The size composition of shipments can vary according to market demand and to enforcement of the regulations (Dr J. Porter, 21 June 2000). Canada Canada regulates its swordfish fishery through input and output controls (DFO 1997). It first restricted entry to its swordfish fishery in 1979 and has since limited the number of longliners to 77 (Folsom and Crory 1997, p. 23). Canada s fishery is seasonal. It uses area closures on the Scotian Shelf during September December to protect swordfish of spawning age. To reduce the bluefin tuna bycatch and catch of small swordfish, it does allow fishing to commence before 1 August west of 65º30 E. Spain Spain has limited the number of permits available in the fishery (how many is not clear). Each boat must hold a Permiso Temporal de Pesca (permission for fishing), which specifies its number of fishing days, areas, time allowed and quota (Dr J. Mejuto, 26 April 1999). Bycatch Blue shark (67% by number), followed by swordfish (17%) and shortfin mako (12%) dominated Spain s longline landings in The ratio of shark to swordfish caught by Spain s longliners increases at lower latitudes. The proportion of shark in the catch also varies seasonally: blue shark are most abundant during April June whereas most shortfin mako shark are taken in July September (Buencuerpo et al. 1998, p. 669). The NMFS observer program reported the following catch breakdown for United States longliners in the North Atlantic during : swordfish (27% by number); yellowfin, bigeye and northern bluefin tuna (21%); other tuna (3%); other finfish (22%); shark and ray (20%); turtle, mammal and seabird (<1%). Note that those data are for swordfish and tuna boats combined and that catch composition is likely to vary considerably with targeting. Blue shark and shortfin mako are the most common shark caught (Cramer 1996b). Large numbers of blue shark occur offshore during July September and in coastal waters during October March. Blue shark may comprise the largest portion of the total catch in northern temperate waters. For example, they comprise up to 77 per cent of the catch on the Grand Banks during April June. (Cramer and Adams 1999, p. 11). There are bans on retaining marlin for commercial sale. The observer data indicate that United States s longliners in the North Atlantic caught 39 marine mammal and 544 turtle in All were released alive. Marine mammal included: humpback 76 Bureau of Rural Sciences

43 North Atlantic whale, minke whale, Risso s dolphin, long-finned pilot whale, short-finned pilot whale, common dolphin, Atlantic spotted dolphin, pantropical spotted dolphin, striped dolphin, bottlenose dolphin and harbour porpoise. Loggerhead and leatherback turtle were the most common turtle caught by longliners. They also caught green, hawksbill and Kemp s Ridley turtle. For 1997 observers also reported that longliners caught 18 seabird which died. They released a further 15 seabird alive (NMFS 1999, pp to 3-223). Canada prohibits the finning of shark. The Department of Fisheries and Oceans monitors the landing of shark fins and they may only be sold in proportions that correspond to the number of carcasses. Interaction Longliners targeting swordfish in the north-western Atlantic often have a bycatch of northern bluefin and those targeting northern bluefin or bigeye sometimes have a swordfish bycatch. Fishers equipped with harpoons will take either northern bluefin or swordfish. In tropical waters, swordfish longliners have a bycatch of tuna, such as yellowfin, albacore and bigeye and those targeting tuna or shark sometimes have a swordfish bycatch. Interaction between those groups is not currently a major management concern, but it does complicate quota monitoring and estimating catch levels for stock assessment. A recreational fishery for swordfish developed off Florida when anglers landed two large swordfish in the early 1970s. Recreational activity peaked during The fishery then diminished as catch rates declined, from swordfish per boat in tournaments in the early 1970s to less than 0.09 in later tournaments (Joseph et al. 1994, p. 23). Such declines caused friction between anglers and longliners in the Gulf of Mexico and off Florida. Advice needs Stock structure and the location of boundaries are critical issues for assessment and fishery management of swordfish. Advice is required on the relationship between Mediterranean and Atlantic swordfish and the relationship between the north and South Atlantic stocks (Dr J. Porter, 16 December 1999). Fishery managers will need more detailed information on mixing rates and the origin of swordfish in particular areas. Such information is required for stock assessment and for the application of stock-specific quotas, e.g., the allocation of swordfish catches in the Strait of Gibraltar against North Atlantic or Mediterranean quotas. Even though there were major improvements in the submission of catch data during the late 1990s, ICCAT needs to further improve the quality of information on swordfish catches in the North Atlantic. Some uncertainty remains over catch estimates of non-members, boats flying flags of convenience and bycatch levels in other fisheries. The activities of non-members continue to diminish the effectiveness of regulations. ICCAT also requires information on the effectiveness of management measures. SCRS has shown that the minimum size regulations, for example, have been largely ineffective because longlining has continued in areas where small swordfish are abundant and mortality rates are high for released swordfish. Assessments also require improved Swordfish fisheries 77

44 estimates of the number of undersized swordfish caught and discarded and likely to die. The problem is more in the implementation of the regulation than an uncertainty for stock assessment. Apart from the longline fisheries of the United States, Canada, Spain and, more recently, Japan and Taiwan, there is a paucity of information on bycatch in the North Atlantic swordfish fishery. To make sound decisions on bycatch issues, ICCAT will require information on bycatch levels and assessments of the effects of fishing on vulnerable species and the pelagic ecosystem as a whole. Future prospects Scientists have been assessing the status of North Atlantic swordfish since Those assessments have consistently indicated a declining trend in the stock, forcing ICCAT to recommend increasingly stricter conservation measures. In many instances, non-member and member governments have ignored those recommendations, while other measures, like the minimum size regulation, have been mostly ineffective. In consequence, the North Atlantic swordfish stock continued to decline, at least until the mid-1990s. The 1999 assessment, however, suggests that the decline in North Atlantic swordfish has slowed and that there were strong recruitments of young (age 1) swordfish in 1997 and Improvements in data (coverage, quality and processing) and reduced catches contributed to the better outlook. Nevertheless, this is just one positive assessment in a series of assessments stretching over a decade. The stock remains at historically low levels and rebuilding the stock will require the long-term commitment of all fishing nations. In particular, non-members have the potential to increase catches above sustainable levels and to disrupt the commitment of members to accept and adhere to management measures. Similarly, the large amount of swordfish taken by distant-water longliners as bycatch will continue to create problems for assessment and management. Environmental groups will claim that the pressure they brought to bear on the United States fishing industry and governments through the Give swordfish a break campaign contributed to a reduction in swordfish catches. However, some scientists believe that the campaign disrupted the relationships between scientists, fishers and managers while providing no real benefit to the stock. United States and Canadian fishers, who were implementing ICCAT recommendations on quotas and minimum sizes, suffered from lower market prices and increased imports. Future activities of environmental groups will hinge on their attitude towards the proposition that the North Atlantic swordfish fishery is now being managed sustainably. At the same time, environmental groups in the United States are paying increased attention to bycatch issues in longline fisheries, such as the practice of shark finning. In response to declining catch rates, many swordfish longliners are turning to mixed fishing strategies, targeting tuna or shark, to take advantage of fluctuations in catch rates and market prices. Many other longliners moved out of the North Atlantic into the South Atlantic. With declines in swordfish catch rates now emerging in the South Atlantic, more longliners are likely to be redeployed to other areas, such as the Pacific and Indian oceans. References Anonymous (2000) Industry s pelagic longline plan. Commercial Fisheries News, January. Bannerot, S.P. and Austin, B.C. (1989) Development of a U.S.-based longline fishery for swordfish (Xiphias gladius) in the Caribbean region: Problems and opportunities. In Waugh, G.T. and Goodwin, M.H. (eds) Proceedings of the Thirty-ninth Annual Gulf and Caribbean Fisheries Institute, Hamilton, Bermuda, November Beckett, J.S. (1971) Canadian swordfish longline fishery. International Commission for the Conservation of Atlantic Tunas SCRS report 80/71/36. Not seen, cited by Dr J. Porter (16 December 1999). Beckett, J.S. (1974) Biology of swordfish, Xiphias gladius L., in the northwest Atlantic Ocean. pp In Shomura, R.S. and Williams, F. (eds) Proceedings of the International 78 Bureau of Rural Sciences

45 North Atlantic Billfish Symposium, Kaila Kona, Hawaii, 9 12 August Part 2. Review and contributed papers. NOAA Technical Report NMFS SSRF-675, Berkeley, S.A. (1989) Trends in Atlantic swordfish fisheries. In Stroud, H.R. (ed.) Planning the Future of billfishes: Research and management in the 90s and beyond. Proceedings of the Second International Billfish Symposium, Kailuna Kona, Hawaii, August Part 1, pp Berkeley, S.A. and Waugh, G.T. (1989) Considerations for regional swordfish management. In Waugh, G.T. and Goodwin, M.H. (eds) Proceedings of the Thirty-ninth Annual Gulf and Caribbean Fisheries Institute, Hamilton, Bermuda, November Berkeley, S.A., Irby, E.W. Jr and Jolley, J.W. Jr (1981) Florida s commercial swordfish fishery: longline gear and methods. Florida Sea Grant, Miami. Marine Advisory Bulletin MAP 14. Brewster-Geisz, K. (1997) Country report: United States. In Folsom, W.B., Weidner, D.M. and Wildman, M.R. (eds) (1997) World swordfish fisheries: an analysis of swordfish fisheries, market trends and trade patterns, past-present-future. Volume I North America. United States Department of Commerce, NOAA Technical Memorandum NMFS F/SPO 23. Buencuerpo, V., Rios, S and Moron, J. (1998) Pelagic sharks associated with the swordfish, Xiphias gladius, fishery in the eastern North Atlantic Ocean and the Strait of Gibraltar. Fishery Bulletin 96, pp Caddy, J.F. (1976) A review of some factors relevant to management of swordfish fisheries in the northwest Atlantic. Canada Department Environment, Fisheries and Marine Services Research and Development Technical Report No Not seen, cited by Dr J. Porter (16 December 1999). Campos, J.L., Rolon, M.A. and Munoz-Roure, O. (1987) The emergence of a swordfish fishery around Puerto Rico and the Virgin Islands. In Williams, F. (ed.) Proceedings of the Thirtyeighth Annual Gulf and Caribbean Fisheries Institute, Trois-Islets, Martinique, November Cramer, J. (1996a) Recent trends in the catch of undersized swordfish by the U.S. pelagic longline fishery. Marine Fisheries Review 58(3), pp Cramer, J. (1996b) Species reported caught in the U.S. commercial pelagic longline, gillnet and pair trawl fisheries from 1987 to Miami Laboratory Contribution MIA 95/ Southwest Fisheries Science Center, Miami. Cramer, J. and Adams, H. (1999) Large pelagic logbook newsletter United States Department of Commerce, NOAA Technical Memorandum NMFS SEFSC 220. Cramer, J., Bertolino, A.R. and Scott, G.P. (1995) Estimates of swordfish discarded dead by United States longline vessels since Collected Volume of Scientific Papers 44(3), pp ICCAT, Madrid. DFO (1997) The Canadian Atlantic swordfish (Xiphius gladius) fishery Integrated management plan. Department of Fisheries and Oceans, Canada. Ehrhardt, N.M., Robbins, R.J. and Arocha, F. (1996) Age validation and growth of swordfish, Xiphias gladius, in the northwest Atlantic Ocean. ICCAT Collected volume of scientific papers 45(2), pp FAJ and NIFSF (1998) National report of Japan. Fisheries Agency of Japan and the National Institute of Far Seas Fisheries. In ICCAT report for biennial period, , Part II (1997) Volume 2. ICCAT, Madrid. Folsom, W.B. (1997) World swordfish fisheries: An analysis of swordfish fisheries, market trends and trade patterns, past-present-future. Volume VI Western Europe. United States Department of Commerce, NOAA Technical Memorandum NMFS F/SPO 23. Swordfish fisheries 79

46 Folsom, W.B. and Crory, D.M. (1997) Country Report: Canada. In Folsom, W.B., Weidner, D.M. and Wildman, M.R. (eds) (1997) World swordfish fisheries: An analysis of swordfish fisheries, market trends and trade patterns, past-present-future. Volume I North America. United States Department of Commerce, NOAA Technical Memorandum NMFS F/SPO 23. Gibson, C.D. (1998) The broadbill swordfishery of the Northwest Atlantic: An economic and natural history. Ensign Press, Maine. Goode, G.B. (1883) Materials for a history of the sword-fishes. Report on U.S. Commercial Fish [1880]8, pp Hoey, J.J. (1995) Bycatch in western Atlantic pelagic longline fisheries. In Solving bycatch: considerations for today and tomorrow. Proceedings of the Solving Bycatch Workshop. Seattle, Washington, September Hoey, J.J. and Casey, J. (1988) Review of the US fishery for swordfish, Collected volume of scientific papers 27, pp Hoey, J.J. and Mejuto, J. (1991) Swordfish size composition data from Spanish and United States North Atlantic longline fisheries. Collected volume of scientific papers 35(2), pp Hoey, J.J., Bertolino, A.R., Cramer, J. and Rogers, C.W. (1995) Recent trends in the US Atlantic longline fishery. Collected volume of scientific papers 44(3), pp Hoey, J.J., Conser, R.J. and Bertolino, A.R. (1989) The western North Atlantic swordfish. Audubon wildlife report, 1989/1990, pp Hurley, P.C.F. and Iles, T.D. (1981) A review of the Canadian swordfish fishery. International Commission for the Conservation of Atlantic Tunas Collected volume of scientific papers 15(1), pp Not seen, cited by Dr J. Porter (16 December 1999). ICCAT (1989) Second ICCAT Swordfish Workshop (Madrid, September ). Collected volume of scientific papers 28, pp ICCAT (1995) Report of the Standing Committee on Research and Statistics (SCRS) Atlantic Swordfish. ICCAT (1997a) Report of the ICCAT Swordfish Stock Assessment Session (Halifax, Nova Scotia, Canada October ): 1996 Detailed report for swordfish. Collected Volume of scientific papers 46(3), pp ICCAT, Madrid. ICCAT (1997b) Report of the Standing Committee on Research and Statistics (SCRS; Halifax, Nova Scotia, Canada October ) Atlantic Swordfish. ICCAT, Madrid. ICCAT (1999a) ICCAT management recommendations & resolutions concerning Atlantic swordfish (Xiphias gladius). < (accessed on 21 April 1999). ICCAT (1999b) 1999 Swordfish executive summary. SWO ATL Atlantic swordfish. ICCAT, Madrid. ICCAT (1999c) 1999 Detailed report for swordfish. ICCAT SCRS Swordfish Stock Assessment Session, Madrid, Spain September 27 to October ICCAT, Madrid. ICCAT (1999d) Report of the Standing Committee on Research and Statistics. Swordfish (Xiphias gladius) Executive summary. International Commission for the Conservation of Atlantic Tunas. < (accessed on 4 February 2000). ICCAT (in press b) Report of the Standing Committee on Research and Statistics (SCRS; Madrid, October 19 to November ) Atlantic Swordfish. ICCAT, Madrid. ICCAT (in press a) 1998 Swordfish Detailed Report. ICCAT, Madrid. Joseph, J., Bayliff, W.H. and Hinton, M.G. (1994) A review of information on the biology, fisheries, marketing and utilization, fishing regulations and stock assessment of swordfish, 80 Bureau of Rural Sciences

47 North Atlantic Xiphias gladius, in the Pacific Ocean. Internal Report No. 24. Inter-American Tropical Tuna Commission, La Jolla, California. Mace, P.M. (1995) An evaluation of the effectiveness of the current minimum size for Atlantic swordfish. Collected volume of scientific papers 44(3), pp ICCAT, Madrid. Mejuto, J., de la Serna J.M. and García, B. (1998) Some considerations on the spatial and temporal variability in the sex-ratio at size of the swordfish (Xiphias gladius). International Commission for the Conservation of Atlantic Tunas, Collected Volume of Scientific Papers, 48(1), pp Mejuto, J., García, B., de la Serna, J.M. (1997) Activity of the Spanish surface longline fleet catching swordfish (Xiphias gladius) in the Atlantic, years 1994 and Collected Volume of Scientific Papers 46. SCRS/96/138. ICCAT, Madrid. Miyake, P. and Rey, J.C. (1989) Status of Atlantic Broadbill Swordfish Stocks. In Stroud, R.H. (ed.) Planning the future of billfishes: research & management in the 90s and beyond. Proceedings of the Second International Billfish Symposium, Kailua Kona, Hawaii, August NMFS (1998) National Report of the United States. In ICCAT report for biennial period, , Part II (1997) Volume 2. ICCAT, Madrid. NMFS (1999) Final fishery management plan for Atlantic tuna, swordfish and sharks. United States Department of Commerce, NOAA, National Marine Fisheries Service, Highly Migratory Species Management Division, Silver Springs. NOAA (1997) Fisheries service to close swordfish fishery to meet ICCAT quota. Media Release NOAA 97-R118. < (accessed on 21 April 1999). NRDC (1998) Swordfish splash. The Amicus Journal, 19(4), p. 3. Ortiz, M., Restrepo V.R. and Turner, S.C. (2000) North Atlantic swordfish sex-ratios at size keys: Analyses and development. International Commission for the Conservation of Atlantic Tunas, Collected Volume of Scientific Papers, xx xx. Madrid. Pickard, G.L. (1979) Descriptive physical oceanography. 3rd (SI) edition. Pergamon Press, Oxford. Podesta, G.P., Browder, J.A. and Hoey, J.J. (1993) Exploring the association between swordfish catch rates and thermal fronts on U.S. longline grounds in the western North Atlantic. Continental Shelf Research 13(2/3), pp Porter, J.M. (1997) Perspective on Atlantic (and Mediterranean) swordfish fisheries and assessments: the ICCAT experience (working document). Paper presented at the Second International Pacific Swordfish Symposium, Kahuku, Hawaii, 3 6 March Porter, J.M. and Allen, C.J. (1998) National report of Canada, Collected Volume of Scientific Papers (in press). ICCAT, Madrid. Restrepo, V.R. (1998) An introduction to 25 years of ICCAT stock assessments. Collected volume of scientific papers 50(5). ICCAT, Madrid. Rey, J.C., Mejuto, J. and Iglesias, S. (1988) Evolucion historica y situacion actual de la pesqueria Esanola de pez espada (Xiphias gladius). Collected volume of scientific papers 38, pp Safina, C. (1998) Song for the blue ocean. Henry Holt and Company, New York. SAFMC (1985) Fishery management plan, regulatory impact review, initial regulatory flexibility analysis and final environmental impact statement for Atlantic swordfish. Prepared by the South Atlantic Fishery Management Council (SAFMC), in cooperation with Caribbean FMC, Gulf of Mexico FMC, Mid-Atlantic FMC and New England FMC. Swordfish fisheries 81

48 Stone, H.H. and Porter, J.M. (1998) Updated age-specific CPUE for Canadian swordfish longline, Collected volume of scientific papers 48(1), pp ICCAT, Madrid. Stone, H.H. and Porter, J.M. (1999) Updated age-specific CPUE for Canadian swordfish longline ( ), with information on nominal CPUE for yellowfin, bigeye and albacore tuna bycatch. International Commission for the Conservation of Atlantic Tunas. Collected volume of scientific papers 49(1), pp Not seen, cited by Dr J. Porter (16 December 1999). Tibbo, S.N., Day, L.R. and Doucet, W.F. (1961) The swordfish (Xiphias gladius), its life history and economic importance in the northwest Atlantic. Bulletin of the Fisheries Research Board of Canada. Volume 130. Uosaki, K. and Uozumi, Y. (1993) The trend of mean length of Atlantic swordfish from 1975 to 1990 caught by the Japanese longline fishery. Collected Volume of Scientific Papers 40(1), pp ICCAT, Madrid. Weidner, D.M., Arocha, F., Fontes, F.J., Folsom, W.B. and Serrano, J.A. (1999) World swordfish fisheries: An analysis of swordfish fisheries, market trends and trade patterns, past-presentfuture. Volume IV Latin America. United States Department of Commerce, NOAA Technical Memorandum NMFS F/SPO Bureau of Rural Sciences

49 Chile Chile s swordfish fishery Overview Chile has what are probably the world s most productive pelagic waters. Productivity is driven by coastal upwellings and the cold, nutrient-rich Peru Current that flows northwards close to Chile s coastline. Chile s jack mackerel resources would seem to offer swordfish an abundant food source, although squid often seem to be the preferred prey of swordfish. Swordfish fishing is highly seasonal and distributed over a wide latitudinal range (18 40ºS). Chile s fishery has two components: subsistence or artisanal fishers using driftnet (and, in the past, harpoon) and commercial longliners. Longliners operate in offshore waters searching for cool (18 19ºC) sea surface temperatures. Small (~75 kg) swordfish characterise Chile s longline fishery. With the introduction of driftnets and large, commercial longliners the fishery rapidly expanded, landing about 7255 t in Within three years of that peak, however, the inshore fishery collapsed. Artisanal fishers have had to travel further offshore to catch swordfish. It is not clear whether the collapse was the result of overfishing or whether it was due to other factors, such as long-term climatic shifts or massive increases in Chile s jack mackerel catch. Swordfish catches recovered to t per year during , with artisanal fishers reporting more than 1800 t in The following is essentially a summary of Weidner and Serrano (1997) who provide a comprehensive description of swordfish fisheries in the south-eastern Pacific Ocean. The fishery Development Archaeological records show that, before the Spanish conquest of the 1500s, the Incas caught and consumed swordfish in what is now Chile (Weidner and Serrano 1997, p. 569; Dr. M. Hinton, 13 November 1999). Since the 1900s artisanal 1 fishers are known to have harpooned swordfish in Chile s coastal waters (Map 4). Initially those were subsistence fishing activities. Artisanal operations expanded (e.g., landing 2146 t in 1946) in response to the arrival of United States freezer boats and the establishment of freezer plants in Chile during the 1930s and 1940s. The fishery then declined as United States domestic production increased and during the 1970s mercury restrictions reduced swordfish imports by the United States. Artisanal fishers were attracted away to other fisheries or to paid employment in other industries (Weidner and Serrano 1997, pp. 429, 572). Experiments with drift gillnet fishing gear in and trial shipments of fresh swordfish to markets in the United States led to a rapid increase in Chile s swordfish catches (Weidner and Serrano 1997, p. 429). With government support for the purchase of fishing boats and equipment, artisanal fishers began using driftnets to catch swordfish in the mid-1980s. Few now use harpoons (Barbieri et al. 1998, p. 2). In the late 1980s commercial fishers acquired large, ocean-going boats and, in 1989, began using longlines to catch swordfish in northern offshore waters. Catches by artisanal driftnet and commercial longline fishers peaked at 7255 t in 1991, but then declined. Conversion of freezer longliner operations to land higher-priced, fresh swordfish reinvigorated the fishery s commercial sector, and catches in offshore waters improved to t per year during (Weidner and Serrano 1997, pp. 430, 467, 529). The artisanal sector eventually recovered too and artisanal fishers reported more than 1800 t in 1998 (Dr R. Serra, 30 October 1999). 1 Originally the artisanal fishing activities were true subsistence operations. Since the 1930s most artisanal fishers have sold their catch and could thus be considered small-scale, commercial fishing operations. However, the term artisanal is retained under Chile s management arrangements to refer to fishing boats smaller than 18 m and 50 GRT; commercial boats are larger than 18 m or 50 GRT. Swordfish fisheries 83

50 Catches Catch trends During artisanal fishers landed about 13 to 455 t of swordfish per year. With the introduction of driftnets and large, commercial longliners the fishery rapidly expanded, landing about 7255 t in 1991 (Weidner and Serrano 1997, p. 429; Figure 15). Expansion coincided with access to international markets for swordfish and the availability of remotely sensed sea surface temperature imagery (Barbieri et al. 1998, p. 2; Yañez et al. 1997, p. 1). Total effort peaked two years later, in Artisanal driftnet effort underwent a tenfold expansion, from an estimated 4777 sea days in 1987 to sea days in 1993, with the total area of driftnet increasing from 0.6 million m 2 to 100 million m 2 in 1991 (Weidner and Serrano 1997, p. 761). Catches declined after 1991 with only 2594 t of swordfish reported in Many commercial longliners left the fishery and artisanal driftnet fishers landed 80 per cent of the catch. With renewed interest in fresh exports and high catch rates in offshore waters by commercial longliners, the catch revived to 3145 t in The increase in catch was in sharp contrast to the decreasing number of artisanal boats involved in the swordfish fishery. Artisanal driftnet PACIFIC OCEAN South East Trade Wind Drift Ju an Fernandez Islands Southern West Wind Drift Valparaiso Santiago San Antonio Talcahuano fishers reported increasing problems finding swordfish in coastal waters and many withdrew from the fishery in 1996 (Weidner and Serrano 1997, pp. 430, 467, 527). Only 133 artisanal boats operated in Size composition The size composition of swordfish catches depends on fishing method and it varies geographically. Worldwide, most of the large swordfish taken by anglers have been taken in Chile, with a world record 536 kg swordfish boated there. In contrast, small swordfish characterise Chile s longline fishery. When they began fishing in northern offshore waters (20 31ºS), longliners reported substantial catches of small (<50 kg), mostly juvenile swordfish. The small size of swordfish often do not justify commercial operations in northern waters (Weidner and Serrano 1997, pp. 482, 496) Arica Coquimbo Map 4. Chile s exclusive economic zone showing major ocean circulation systems (Weidner and Serrano 1997, p. 508), bathymetry and main swordfish fishing ports. White indicates ocean depths of less than 1000 m, light grey indicates depths of m and dark grey indicates depths greater than 2000 m. Also shown is Chile s capital city, Santiago. The northern border of Chile s exclusive economic zone (dashed line) is approximate. Peru Current Bureau of Rural Sciences

51 Chile 8 Longline Artisanal Catch ('000 t) Year 0 Figure 15. Annual catches of swordfish reported by Chile s longliners and artisanal fishers (FAO 1997; data from Dr R. Serra, 30 October 1999). The largest swordfish tend to be taken in coastal waters, although this is at least partly due to the driftnet and harpoon fishing methods used there. The large-mesh driftnets do not entangle small swordfish, whereas baited longline hooks are much less selective. In an area off central Chile where driftnets and longlines are used on the same grounds, swordfish taken by driftnet average 122 kg T wt (~165 kg W wt ) compared with 75 kg T wt (~100 kg W wt ) for longline-caught swordfish. A small sample of swordfish caught by harpoon averaged 152 kg T wt (~205 kg W wt ) in 1987 and 1988 (Weidner and Serrano 1997, pp , 755). The driftnets appear to take a much larger proportion of female swordfish, whereas there are insufficient data to indicate the sex composition of longline catches (Weidner and Serrano 1997, p. 492). Species associations Shark are the most common bycatch in Chile s longline fishery. Mako shark predominate, but blue shark are reported too. The incidental catch of tuna, such as yellowfin, bigeye and albacore, is small (<5% of the weight of longline catches). Tuna bycatch is rare because of the cool (18 19ºC) waters fished and the shallow (35 55 m maximum depth) longlines that are used (Weidner and Serrano 1997, pp ). Distribution Swordfish occur off Chile in both coastal and offshore waters, from 18 40ºS. For years distantwater longliners, mostly from Japan, had fished for tuna at latitudes similar to but generally outside what is now Chile s 200 nautical mile exclusive economic zone. Spain s longliners also fish for swordfish in the area. However, there is uncertainty over 2 Cited in Weidner and Serrano (1997, p. 776). Swordfish fisheries 85

52 the distribution of swordfish over the temperate waters of the wider south-eastern Pacific because longlining, especially for swordfish, has been rare there (Weidner and Serrano 1997, pp. 429, 464, 466). In the 1940s and 1950s the artisanal harpoon fishery was based in northern ports (~20ºS). They were most successful in green water during the relatively calm months of January May. The introduction of driftnets extended the fishing season to July and sometimes September and broadened the area fished. Most of the artisanal driftnet boats were based in central ports (~35ºS) and operated within 35 nm of the coastline during the 1980s. The reason for the southwards shift in activity is not well understood. It might have been due to harpoons not being as effective in the southern waters as driftnets, or the availability of larger boats capable of operating in southern waters or it might have been related to a change in the distribution of swordfish (Weidner and Serrano 1997, pp. 549, ). Following the 1991 peak in catches, artisanal driftnet and commercial longline fishers report having to progressively move further offshore to maintain the economic viability of their fishing operations. Artisanal driftnet fishers now fish from central ports (31 42ºS) out to about 100 nm from the coastline (Weidner and Serrano 1997, p. 467, 513). Longliners operate off northern Chile (19 33ºS) and, because of government regulations, outside of 120 nm. During the early 1990s Chile s longliners reported limited activity from outside the exclusive economic zone (e.g., t or 9 21% of their total catch). There was a sudden change in the fishery s distribution in 1996 with 72 per cent (2274 t) of the swordfish catch taken outside Chile s zone. Most of that activity was nm offshore and further north (~20ºS) than in previous years (Weidner and Serrano 1997, pp. 467, , 777). Chile s fishery is highly seasonal. Foreign longline data suggest a circular pattern to the movement of swordfish that corresponds to the prevailing surface currents of the south-eastern Pacific. Swordfish follow the northward flowing Peru Current 3 along the South American coastline during July September, then spread out westwards to ºW in October December, returning in January March to Chile s coastal waters, sometimes as far south as 45ºS. They are caught along the central coastline during January March (Weidner and Serrano 1997, p. 474). Artisanal harpoon fishers based in central ports usually report the first catches of the year and some reports suggest that larger swordfish are the first to arrive. Fishing activity then shifts northwards, perhaps as warm water recedes. Although varying from year to year, catches usually peak in April June, then decline during July September (Weidner and Serrano 1997, p. 471). The duration of the longline season has gradually expanded, from March June in the late 1980s extending to September in more recent years. Further offshore, in international waters, the longline season tends to start later, continuing until about December, with best catches in May August. Several companies operate longliners year-round and small quantities of swordfish are caught by driftnet throughout the year (Weidner and Serrano 1997, p. 482). Oceanography Many islands occur within Chile s exclusive economic zone, connected by oceanic ridges. Chile s continental shelf is very narrow, ranging from about 45 nm wide in the far south to 8 nm in the north (Weidner and Serrano 1997, p. 507). Following is a brief description of the four oceanographic features that influence waters where Chilean fishers target swordfish. A westerly flow of surface waters at mid-latitudes in the South Pacific Gyre. Circling the Antarctic, the nutrient-rich West Wind Drift splits as it approaches southern Chile at about 50ºS. The southern branch flows south-east around Cape Horn while the northern branch becomes the Peru Current. 3 The Peru Current is also known as the Humboldt Current. 86 Bureau of Rural Sciences

53 Chile The Peru Current flows northwards along the coastline of Chile and Peru. Off northern Chile the current is about 13 nm wide and located nm offshore. Its flow is particularly strong at m below the sea surface. It is several degrees cooler than adjacent oceanic waters, creating strong temperature fronts and eddies. Along the coastline of Chile and Peru, southerly and south-easterly winds push surface waters away from the coastline drawing up to the surface cold, nutrient-rich water from several hundred metres below. This upwelling lowers the temperatures of coastal waters by a further 2 4ºC. Upwelling and the Peru Current involve water that is several degrees cooler than surrounding oceanic areas. Nevertheless, waters where temperatures are most suitable for swordfish fishing (16 19ºC) are common off central and northern Chile. As a consequence of upwelling and the Peru Current, Chile s coastal waters are considered some of the richest, most productive fishing areas in the world. (Weidner and Serrano 1997, p ). Well known are the effects of El Niño warming events on the Peruvian anchovy (sardine) fishery. For example, Peru s anchovy catch fell by more than 80 per cent, from 12.5 million tonne in 1970 to 2.3 million tonne in 1972 as a result of an El Niño event. In Chile, however, El Niño events have not had such severe effects on fisheries for small pelagic fishes (e.g., during Chile s anchovy catch fell from 1.4 million tonne to 1.2 million tonne). (Weidner and Serrano 1997, p. 510). Gear and targeting Fleets During the 1930s and 1940s United States and other sport fishers targeted swordfish in similar areas to the artisanal harpoon fishery. In the 1950s recreational fishing declined for unknown reasons. Currently recreational fishers infrequently target swordfish in Chile (Weidner and Serrano 1997, p. 470). The average size of artisanal boats increased during the 1980s and 1990s. In 1996 they averaged 15 m and most are GRT. Many are wooden, but new boats are often steel and as large as government regulations permit (18 m and 50 GRT; Weidner and Serrano 1997, pp. 533, ; Barbieri et al. 1998, p. 3). Artisanal boats are categorised as three types: small (7 9 m) open-decked bongas ; less common now, slightly larger (9 12 m) faluchos, which are traditional boats that used harpoons; and lanchas, which deploy driftnets and are 9 18 m (Weidner and Serrano 1997, p ). Most artisanal boats are multipurpose, rather than purpose-built. Some have hydraulic winches for retrieving the net, but many manually haul the net. Initially, trips lasted about three days, but with larger boats and the need to operate further offshore, artisanal trips of 12 days are more common. They have six to eight crew members (Weidner and Serrano 1997, pp. 531, ). Swordfish fisheries 87

54 Estimates vary of the number of artisanal driftnet boats targeting swordfish during the 1980s and 1990s. Instituto de Fomento Pesquero (IFOP) data suggest that about 25 artisanal boats fished in 1986, increasing to about 450 boats by 1989 and peaking at 500 in 1991 when Servicio National de Pesca (SERNAP or SERNAPESCA) closed the fishery to new entrants (Figure 17). Other reports suggest that there may have been as many as 1000 boats involved in 1991, although many of those had an opportunistic involvement in the swordfish fishery (Weidner and Serrano 1997, pp. 526, 578). Chile has a substantial longline fleet, but most boats are currently involved in demersal longlining for species such as Patagonian toothfish in Antarctic waters. A few of Chile s pelagic longliners are based in Montevideo and target swordfish in the South Atlantic. Those longliners also fish for toothfish at times (Weidner and Serrano 1997, p. 525). Up to 143 commercial longliners fished for swordfish during Most are purpose-built. They range in size up to 92 m, but many are small or medium-sized (18 26 m). Most are steel-hulled. Declining El Niño Southern Oscillation (ENSO) Broad-scale oceanographic events, like ENSO, influence the distribution and abundance of pelagic fish by their effects on water temperature, productivity and ocean circulation. ENSO events are driven by the Walker Circulation which involves the trade winds bringing moist air to the Indonesia region. There the air rises, creating rainfall and an area of low pressure. The air then travels eastwards at high altitudes before sinking over the eastern Pacific where it creates dry conditions and high air pressure. The Southern Oscillation Index (SOI) is a measure of the strength of the Walker Circulation. It is the difference in air pressure between Tahiti and Darwin. A negative index indicates an El Niño event whereas a positive index indicates a La Niña event. An El Niño event involves a weakening of the Walker Circulation, causing the equatorial west Pacific to cool, trade winds to ease and rainfall to decrease over Australasia. A La Niña event involves a strong Walker Circulation, resulting in warm water in the equatorial west Pacific, strong trade winds and increased rainfall over Australasia. In the Atlantic, broad-scale oceanographic events may be predicted by the North Atlantic Oscillation (NAO), which is associated with changes in the strength and pathways of winter storms crossing the Atlantic from North America to Europe (Visbeck et al. 1999, p. 1). prices for frozen swordfish after 1992 resulted in many longliners withdrawing from the swordfish fishery (Figure 16). Many longliners are freezer boats that are now being used for shorter trips where swordfish are stored fresh, on ice. They usually carry 15 or 16 crew members (Weidner and Serrano 1997). These are true, distant-water operations, with one longliner reportedly fishing 700 nm offshore, at 85ºW. The trips of freezer longliners were up to two months in duration, but with the shift to fresh operations most trips are no more than 15 days. They undertake nine or ten trips per year (Weidner and Serrano 1997, pp. 521, 558). Fish-finding Research during the 1960s suggested that 17 19ºC was the optimal range of sea surface temperature for harpooning swordfish in Chile (Tobella 1964). 4 Artisanal driftnetters usually fish in slightly cooler (16 17ºC) coastal waters. In contrast, longliners, which operate in warmer offshore areas, search for sea surface temperatures of 18 19ºC or warmer. Swordfish are most abundant outside areas of upwelling and along the oceanic margins of the Peru Current, but within mixing areas with weak temperature gradients ( C per nm; Weidner and Serrano 1997, p. 568). 4 Cited in Weidner and Serrano (1997, p. 476). 88 Bureau of Rural Sciences

55 Chile Year 0 Figure 16. Estimates of annual fishing effort of Chile s artisanal fleet ( : Barbieri et al. 1994; : Yáñez and Toro 1999b, p. 10). 600 Longline Artisanal * Year *no estimate available for artisanal fleet Figure 17. Estimates of the annual numbers of boats in Chile fishing for swordfish (Artisanal: Babieri et al. 1997; longline: SERNAP 1997). 6 5 Cited in Weidner and Serrano (1997, p. 764). 6 Cited in Weidner and Serrano (1997, pp. 741, 743). Swordfish fisheries 89

56 As early as the 1930s artisanal harpoon fishers reportedly searched for differences in ocean colour. Driftnet and longline fishers often search for transition zones, where nutrient-rich water of the Peru Current mixes with clear oceanic water. The ideal water colour for swordfish fishing may vary seasonally too, with best catches in light blue water in January March and in greenish water in April May (Weidner and Serrano 1997, p. 490). Following seabird or schools of jack mackerel produces good swordfish catches for some fishers. Others are reported to search for oil slicks created by defecating swordfish (Weidner and Serrano 1997, pp , 552). Fishing gear In the mid-1980s artisanal fishers acquired monofilament driftnets. Most driftnets are about km long and 55 m deep (Barbieri et al. 1998, p. 4). They are comprised of very large (51 56 cm stretched) mesh of mm polypropylene with a lead line along the net bottom and buoys every 45 m. The driftnets are suspended 9 60 m below the sea surface and may reach down to a depth of 120 m. Since 1985 artisanal fishers have attached lightsticks every 40 m along their driftnets (Weidner and Serrano 1997, pp ). Chile s longliners use two types of gear: American-style longlines and, more commonly, Spanishstyle longlines (Table 7). Some artisanal boats also use longlines of about 1000 hooks per set (Weidner and Serrano 1997, p. 545). Spanish-style longlines consist of monofilament nylon mainlines, monofilament nylon branchlines and galvanised wire leaders. The boats have two hydraulically powered plates that revolve to haul the longline (Weidner and Serrano 1997, p. 545). The hooks of Spanish-style longlines lie at m below the sea surface. The American-style longlines involve fewer hooks ( hooks per set), set Table 7. Comparison of features of American-style and the more common Spanish-style longline gear used by Chile s longliners. Longline characteristic Spanish-style American-style depth of deepest hook ~33 55 m ~90 m leader material galvanised wire polypropylene bait horse mackerel, squid and, occasionally, jack mackerel squid deeper (~90 m for the deepest hooks) with more space between each hook (most longlines are km long). The American-style longliners retrieve the mainline with a hydraulically powered drum and use polypropylene leaders. All longliners use lightsticks and many fishers attach a lightstick to every branchline. (Weidner and Serrano 1997, pp. 545, 560). In offshore waters, Spanish-style longlines use horse mackerel and squid for bait, which are the prey of swordfish in those waters. The American-style longlines use squid for bait. Swordfish in coastal waters often feed on jack mackerel and are targeted with fishing methods (harpoon and driftnet) that do not require bait. (Weidner and Serrano 1997, p. 560). Most artisanal and longline fishers use satellite imagery to monitor sea surface temperatures (previously they had to expend considerable fuel and time searching for optimum fishing conditions (Weidner and Serrano 1997, pp , 566). Universities and commercial companies provide sea surface temperature maps. Analyses of remotely 90 Bureau of Rural Sciences

57 Chile sensed data also suggest a link between swordfish catches and high concentrations of phytoplankton (Barbieri and Yañez 1997, p. 7). Most longliners have radar, satellite navigation, video plotters and colour sounders (Weidner and Serrano 1997, p. 536). Fishing practices Driftnets are usually deployed at dusk and hauled at dawn. Longliners usually set at sunset. It takes 5 7 hours to deploy longlines, with 5 8 hours soak time and hours for retrieval (Weidner and Serrano 1997, pp. 537, 559, 583). Discarding of target species Weidner and Serrano (1997) make no mention of Chilean fishers discarding small or damaged swordfish. Longliners have reported dolphin robbing baits from their lines, but killer whale are rarely a problem (Weidner and Serrano 1997, pp ). Shark damage must be common. Since the 1997 introduction of a minimum size limit, longliners are likely to have had to discard large numbers of small swordfish that they catch. Mako shark are retained, whereas the fins are removed from other shark species and the carcass discarded (Weidner and Serrano 1997, pp ). Weidner and Serrano (1997) do not indicate the fate of other pelagic fishes; we presume that most sashimi tuna (yellowfin and bigeye) are airfreighted, along with swordfish, to the United States. Other species are probably retained for consumption by crew or for barter or sale. Assessment Data collection Few data were collected from the artisanal harpoon fishery, despite its long history. Chile s data on the artisanal fishery are primarily port landings, although the Instituto de Fomento Pesquero began collecting data by area (Weidner and Serrano 1997, pp. 448, 465). In the early 1990s the Sub-Secretaría de Pesca (SSP) began collecting detailed data on the swordfish fishery. Swordfish boats are required to provide Servicio National de Pesca with data on their fishing activities and to carry an observer if requested (Weidner and Serrano 1997, pp. 448, 646). However, there is as yet no daily logbook for longliners (or artisanal boats; Dr R. Serra, 30 October 1999). There is no observer program in place and obtaining a berth on artisanal and commercial boats is a common problem for Chilean researchers (Weidner and Serrano 1997, p. 649). Research The Instituto de Fomento Pesquero, a semi-autonomous unit of the Sub-Secretaría de Pesca, conducts or coordinates several research projects relevant to the swordfish fishery. Commercial species, such as swordfish, are the focus of research and scant attention is given to bycatch (Weidner and Serrano 1997, pp ). Projects include fishery data collection and studies of fishery distribution and gear, age and growth, feeding, reproduction and environmental effects on fishing success. The environmental work involves longline catch rates related to remotely sensed (satellite) data on sea surface temperature (Dr R. Humphreys, 23 October 1999). Stock assessment Fishery indicators After swordfish catches peaked in 1991 total catches declined to less than half the peak level. Skill and efficiency are likely to have improved in the longline fishery (Weidner and Serrano 1997, p. 541), yet catch rates declined with increased effort and catches during the 1990s. There are suggestions that the size of swordfish has been declining, although this might be confounded by the shift to offshore waters where swordfish tend to be smaller and by size limits introduced in 1997 (Weidner and Serrano 1997, pp. 496, 504). Artisanal catch rates declined from 391 kg/sea day in 1987 to 89 kg/sea day in 1991, the year of peak total catches (Barbieri et al. 1998, p. 6). 7 Barbieri et al. (1998, p. 6) attributed the decline to 7 Barbieri et al. (1998) do not indicate whether the catch rates are in terms of trunk weight or whole weight. Swordfish fisheries 91

58 changes in the distribution or availability or abundance of swordfish. Anecdotal reports suggest that fishers are undertaking progressively longer trips to maintain the viability of their operations, with artisanal fishers now ranging nm offshore. Artisanal fishers reported increased catch rates in 1996, coinciding with decreased longline and artisanal effort (Weidner and Serrano 1997, p. 504). Stock assessment models In the late 1980s the Inter-American Tropical Tuna Commission (IATTC) had applied a Deriso Schnute delay-difference model to Japan s and Mexico s longline data for the eastern Pacific Ocean. Although Japan s swordfish catch rates declined with increasing effort, the Inter- American Tropical Tuna Commission s assessment was inconclusive because of a lack of contrast in the data. In 1999 preliminary work that added data to the model indicated that catch rates had continued to decline (IATTC 1999, p. 4). From a virtual population analysis, Barbieri et al. (1998, pp. 5 7) estimate that swordfish biomass in Chile fell by about 50 per cent between 1987 and A surplus production model estimated a maximum sustainable yield of about 5800 t at an optimal effort of about artisanal fishing days. The production model suggested that the resource was exploited at optimum levels during 1989, 1991 and Abundance then declined as effort exceeded the optimum. Barbieri et al. s assessments can be considered preliminary because of a lack of information on some parameters, such as growth rate and natural mortality and assumptions about swordfish in Chile comprising a unit stock and the relationship between catch rates and abundance. In 1999 the Universidad Catolica de Valparaiso applied production models (ASP and ADAPT) to artisanal and longline data. Those preliminary analyses included a term for sea surface temperature effects. The models predicted maximum sustainable yields of about t and optimal effort of about fishing days. They also showed an 80 per cent decline in the biomass of swordfish in Chile s exclusive economic zone between 1987 and 1998 (Yáñez and Toro 1999a, 1999b pp. 4, 10, 13). Status Declining catch rates during the mid-1990s might suggest that swordfish have been fished above the level that the swordfish resource in Chile s exclusive economic zone could support. The long time series of artisanal catches suggests that it might only support an annual constant catch of about 900 t, which is the average artisanal catch during Other factors, such as increased catches of jack mackerel (from two million tonne in 1986 to five million tonne in 1995) and broad-scale oceanographic events, may also explain the reduced abundance of swordfish in Chile s exclusive economic zone (Weidner and Serrano 1997, pp. 503, 506). Reliability of the assessment For Chile s exclusive economic zone, there is no reliable quantitative assessment of the status of swordfish and fishery indicators are open to various interpretations. Weidner and Serrano (1997, pp ) provide a list of possible explanations of the decline in swordfish catches and the collapse of the artisanal fishery during the mid-1990s. The following builds on their list: ageing fleet (many artisanal boats had become old and expensive to operate and were withdrawn from the fishery and not replaced); reduced government incentives (the government reduced and eventually removed subsidies for artisanal fishers); poor exchange rates (a stronger Chilean peso had reduced profitability); growth overfishing (commercial longliners had taken excessive quantities of juvenile swordfish that did not then recruit to the adult population exploited by artisanal fishers); However, comparison of annual landings of the artisanal fleet with catch rates multiplied by sea days suggests that the catch rates are in terms of whole weight. 92 Bureau of Rural Sciences

59 Chile recruitment overfishing (artisanal fishers had affected on the adult swordfish population, resulting in reduced recruitment of juvenile swordfish to the fishery); foreign fishing (Activities of Japan s and Spain s longliners outside Chile s exclusive economic zone had adversely affected the swordfish stock. Note, however, that the catch levels of swordfish reported by Japan s longliners in the south-eastern Pacific are low compared with Chile s catch; Weidner and Serrano 1997, p. 505); environmental conditions (e.g., abnormally warm water during the early 1990s (Yañez et al. 1997, p. 1) associated with an El Niño event may have affected swordfish recruitment, abundance or distribution); or pressure on forage species (The expansion and large catches of small pelagic species, such as jack mackerel, may have affected swordfish abundance or distribution). However, there is uncertainty over the link between the abundance of small pelagic fish and swordfish. Whereas jack mackerel are a common food, squid seem to be the preferred prey of adult swordfish). For the Inter-American Tropical Tuna Commission s assessment of eastern Pacific swordfish based on Japan s and Mexico s data, catch rates were higher than those corresponding to the average maximum sustainable yield, suggesting that yield and its corresponding optimum effort might have been underestimated (IATTC 1999, p. 4). Those estimates should be used with caution because the history of declining catch rates with increasing fishing effort in the eastern Pacific provides a tenuous data series for modelling the dynamics of the swordfish population (Dr. M. Hinton, 13 November 1999). Japan s share of the eastern Pacific catch has declined from about 80 per cent during the 1970s to less than 32 per cent during the 1990s. Data from other swordfish fisheries (e.g., Chile) should be included in assessments because Japan s data might not be representative of the entire eastern Pacific fishery in later years (IATTC 1999, pp. 3, 8). Due to uncertainties over stock structure and distribution, the Inter-American Tropical Tuna Commission assessment of eastern Pacific swordfish considered open and closed population models. Its assessment is considered preliminary (IATTC 1999, p. 4). Hinton and Deriso (1998) suggest that there may be two stocks in the eastern Pacific as well as an extension of a northwestern Pacific stock into the area on a seasonal basis. Management Institutions Within the Sub-Secretaría de Pesca, the Servicio National de Pesca collects data and enforces fishery regulations, mainly at landing sites. For the wider eastern Pacific, the Inter-American Tropical Tuna Commission is responsible for tuna, tuna-like species and dolphin. However, Chile is not a member of the commission. Objectives During the 1980s government activities were very successful in promoting the development of Chile s swordfish fishery. Government is now more focused on fishery Swordfish fisheries 93

60 management, arresting the declining yields and declining swordfish size. Resource allocation between artisanal and commercial fishers is another important fishery management issue. Management measures The Sub-Secretaría de Pesca began regulating the fishery in 1990, after the rapid expansion of Chile s swordfish fleet. Their approach was to control fishing effort to prevent overfishing. In 1991 the Sub-Secretaría de Pesca closed the swordfish fishery to new entrants in both the artisanal and commercial sectors. Operators are not permitted to upgrade to larger boats. Artisanal driftnets are limited to a length of 2.47 km and an area of 83.7 km 2. Commercial driftnets are limited in area to km 2 (Weidner and Serrano 1997, p ). The Sub-Secretaría restricted longline fishing effort through gear restrictions and closed the fishery to new entrants. Government regulations prohibit longlining within 120 nm of the coastline, so that coastal waters are reserved for artisanal fishers. They limit the number of hooks that longliners can set, depending on the boat s size. Artisanal longliners are limited to 1200 hooks (Weidner and Serrano 1997, pp. 430, 468, 545, 646). Longliners take many small, juvenile swordfish. On the basis of scientific assessments, in 1997 the Sub-Secretaría de Pesca introduced a minimum size of 58 kg T wt (~80 kg W wt ) for swordfish. A bycatch of 20 per cent juvenile swordfish is permitted and artisanal fishers are allowed to land two small swordfish when they land less than ten swordfish. However, the new regulations are controversial because they are likely to directly affect the viability of the commercial longline fleet. Longline fishers argue that, instead of restrictions on the catch of small swordfish, it is the large, adult swordfish that need protection. The driftnet catch is predominantly dead when landed, so that minimum size restrictions have little effect in reducing mortality of small swordfish in the driftnet sector (Weidner and Serrano 1997, pp ). In international waters, the Sub-Secretaría de Pesca limits longliners smaller than 28 m to 1200 hooks per set. Those larger than 28 m are permitted up to 2000 hooks. The 106 cm minimum size limit also applies to activities in international waters (Dr R. Serra, 30 October 1999). Bycatch Scant information is available on bycatch in Chile s swordfish fishery. Longline sets at shallow depths (35 55 m), like those used by the Spanish-style longliners, tend to catch large numbers of shark. Blue shark are the main bycatch of commercial longliners. Longliners are known to occasionally catch marlin, which are said to be more abundant offshore. Marlin and sailfish bycatch is comparatively rare in the artisanal fishery (Weidner and Serrano 1997, p. 430). Driftnets occasionally take butterfly mackerel. There is concern over incidental catches by driftnets of cetacean, seal (in coastal waters) and turtle, particularly endangered leatherback turtle. Some consider that driftnet fishing was a major cause of a decline in leatherback turtle abundance since the mid-1980s (Weidner and Serrano 1997, pp. 430, ). Interaction Concern over the adverse effects of commercial longlining on the artisanal harpoon and driftnet fisheries resulted in the introduction of substantial area closures and swordfish size restrictions. Chilean fishers are also concerned at the potential impacts on swordfish of foreign longliners (mainly Spain s and Japan s) operating outside their exclusive economic zone. Spanish longliners use ports in Peru, but are not permitted access to Chile s zone or ports (Weidner and Serrano 1997, p. 586). In 2000 Spain presented a case before the World Trade Organization (WTO) against Chile for not allowing Spain s longliners to land swordfish at Chile s ports (Dr R. Roa, 16 May 2000). Regardless, the levels of activity and swordfish catch by longliners adjacent to the zone (e.g., 2000 t in 1997 and much less in earlier years) do not seem to be especially high by international standards. In the early 1950s a few anglers from the United States targeted large swordfish in Chile. However, there is currently little interest in angling there for swordfish or other billfish. 94 Bureau of Rural Sciences

61 Chile Advice needs Government efforts to manage the fishery are complicated by the lack of information on stock structure. An understanding of movement patterns and exchange rates between coastal and offshore waters would help decisions on the allocation of the resource between the commercial and artisanal sectors. A well-planned tagging program could provide useful information on mixing rates, growth and mortality. A comprehensive logbook program, supplemented with observer deployment, would also provide a guide to fishery fluctuations (Dr J. Joseph, 7 May 2000). Elucidation of links between swordfish availability and broad-scale oceanographic events, such as El Niño would be useful for stock assessment and, if it were possible, the prediction of future availability. Similarly, an understanding of the relationship between swordfish and small pelagic fish, such as jack mackerel and the impacts of fisheries for those species on swordfish abundance and distribution would be useful to fishery managers. Future prospects Future prospects for Chile s fishery hinge on swordfish price and their availability in coastal waters (Dr R. Serra, 30 October 1999). Large longliners require substantial initial investments and are expensive to operate. They must relocate or switch to other species when swordfish are not abundant (Weidner and Serrano 1997, p. 529). In contrast, the artisanal fleet is more flexible in the level of fishing activity; fishers have low overheads and can often switch to other fisheries or other industries when swordfish are not abundant. Nevertheless, declining yields have created a downward spiral, so that owners do not have the capital available to replace ageing boats that are costly to operate. Consequently, they continue to operate outdated boats or simply withdraw from the fishery. Chile has a sizeable longline fleet, but most boats are currently involved in demersal longlining in Antarctic waters. If access or catch rates were to decline suddenly in those fisheries, one option might be for those longliners to target swordfish in offshore and more distant waters. References Barbieri, M.A., Canales, C., Correa, V. Donoso, M., Casanga, A.G., Leiva, B., Montiel, A. and Yañez, E. (1998) Development and present state of the swordfish, Xiphias gladius, fishery in Chile. In Barrett, I, Sosa Nishizaki, O. and Bartoo, N. (eds) Biology and fisheries of swordfish, Xiphias gladius: Papers from the International Symposium on Pacific Swordfish, Ensenada, Mexico, December United States Department of Commerce, NOAA Technical Report NMFS 142. Barbieri, M.A. and Yañez, E. (1997) Environmental conditions in the area of the fishery distribution. Paper presented at the Second International Pacific Swordfish Symposium, March , Kahuku, Hawaii. Swordfish fisheries 95

62 Hinton, M.G., and Deriso, R.B. (1998) Distribution and stock assessment of swordfish, Xiphias gladius, in the eastern Pacific Ocean from catch and effort data standardized on biological and environmental parameters. pp In Barrett, I, Sosa Nishizaki, O. and Bartoo, N. (eds) Biology and fisheries of swordfish, Xiphias gladius: Papers from the International Symposium on Pacific Swordfish, Ensenada, Mexico, December United States Department of Commerce, NOAA Technical Report NMFS 142. IATTC (1999) Assessment of swordfish in the eastern Pacific Ocean. Background Paper 7, 63rd meeting of the IATTC. Inter-American Tropical Tuna Commission, La Jolla, California. Visbeck, M., Reverdin, G., Schott, F., Carton, J., Stammer, D., Owens, B. and Clarke, A. (1999) Atlantic Climate Variability (ACVE). < ~visbeck/acve/oceanobs99_acve.html> (accessed on 18 May 2000). Weidner, D.M. and Serrano, J.A. (1997) South America: Pacific. Part A, Section 1 (Segments A and B) in Latin America. World swordfish fisheries. An analysis of swordfish fisheries, markets and trade patterns present and future. Volume IV. NOAA Technical Memorandum NMFS F/SPO 27, NMFS, Silver Spring, Maryland. Yañez, E., Barbieri, M.A., Silva, C. and Nieto, K. (1997) Oceanography and the swordfish fishery in Chile. Paper presented at the Second International Pacific Swordfish Symposium, March , Kahuku, Hawaii. Yáñez, E. and Toro, R. (1999a) Evaluación del stock e influencias del ambiente en la pesquería del pez espada (Xiphias gladius, Linnaeus 1758) en Chile. Informe Final Proyecto DI-UCV N 986/98, Estudios y Documentos N 5/99, Escuela de Ciencias del Mar, Universidad Católica de Valparaíso. Not seen, cited in Yáñez and Toro (1999b). Yáñez E. and Toro, R. (1999b). Monitoreo y evaluación de la pesquería de pez espada en Chile. Paper presented at the United States Chile Swordfish Workshop. Valparaíso, Chile, August Bureau of Rural Sciences

63 Japan's fishery Japan s north-west Pacific swordfish fishery Overview The Japanese have caught pelagic species, such as northern bluefin tuna and swordfish, for more than 100 years in Japan s coastal waters. During longliners began to venture into distant waters where they targeted swordfish and albacore. In the mid-1960s most distant-water longliners installed super-cold freezers and commenced targeting high-priced sashimi tuna. In 1974 many began deep longlining to catch bigeye. For most of Japan s longliners, swordfish are now a bycatch of operations targeting bigeye and yellowfin. Whereas many longliners switched to sashimi tuna in the mid-1960s a small fleet of coastal and offshore longliners continued to target swordfish. We did not access information on the coastal longliner fleet, although some commenced submitting logbooks in Since 1952 Japan has collected catch and effort data from offshore and distant-water longliners. Distributed over a large area of the Pacific, Japan s longline fishery has accounted for a large proportion of the total catch of swordfish, providing a substantial time series of data for stock assessment. Several tentative assessments of the swordfish stocks have been undertaken; they suggest that North Pacific swordfish were not exploited at levels likely to cause a change in catch rates until at least the early 1980s. More recent analyses are inconclusive. However, there is as yet no clear evidence that swordfish are being harvested above levels of maximum sustainable yield. Future assessment will require a better understanding of the stock structure of swordfish in the Pacific and improvements in programs collecting catch, effort and size data. The fishery Development Longlining began in Japan during the Meiji Era ( ). Between 1897 and 1935 the government promoted the development of tuna longlining by granting subsidies to encourage boat building. Developments in engines, line haulers, refrigeration and radios Okhotsk Sea Gyre Oyashio Current Kuroshio Extension North Pacific Current Taiw an Bonin Island s Kuroshio Current Higashioki Fishing Grounds Hawaii North Equatorial Current E 170W 150 Map 5. The north-western Pacific Ocean, showing major ocean circulation systems (Chikuni 1985, pp. 4 6; Roden 1991, p. 2) and the location of the Higashioki Fishing Grounds. Swordfish fisheries 97

64 allowed most of the longliners to fish farther from shore by the mid-1920s. In 1939 Japan s western and central Pacific fleet included 72 longliners of GRT (Wildman 1997, pp ). The Second World War caused substantial declines in the fishery, with catches in 1945 lower than those 50 years earlier. Following the war, Japan s Government again promoted distant-water fishing. Immediate expansion of the fishing grounds was limited initially by a series of boundary lines, known as the MacArthur lines. They allowed for the gradual expansion of fishing activity eastwards to 180 W and southwards to the equator by The last MacArthur line was lifted in 1952, providing Japan s boats with access to international waters around the world. By the mid- 1960s Japan s boats were fishing throughout the northern and south-western Pacific Ocean (Wildman 1997, pp ). Initially, Japan s longliners targeted albacore and swordfish (Sakagawa 1989, p. 67). In Japan, swordfish is occasionally sold as sashimi, but most is marketed as steaks for Teriyaki (Nakamura 1985, p. 51). To target swordfish Japan s longliners used shallow night sets with squid as bait. Japan landed more than swordfish in 1960 compared to about per year in the 1990s (Figure 18). After 1963 longliners changed to mostly day sets, varying baits and gear to target tuna for sashimi markets. Most swordfish were a bycatch to tuna longlining and total catches of swordfish declined (Wildman 1997, pp ). In 1974 offshore longliners fishing the western equatorial Pacific began using deep longline sets to catch high value bigeye (Miyabe and Bayliff 1987, p. 17). The large number of branchlines (>12) between buoys allowed the gear to reach depths of 300 m or more. Deep longline sets caught fewer swordfish and catch rates consequently declined again. There is a directed longline fishery specifically targeting swordfish in the coastal and offshore waters around Japan, in the area traditionally known as the Higashioki fishing Catch (million) Year Figure 18. Annual catches (number) of swordfish reported by Japan s distant-water and offshore longliners in the North Pacific, west of 140ºE (Dr K. Yokawa, 8 June 1999). See Figure 19 for catches (weight) taken by coastal longliners, driftnet boats and harpoon boats in the North Pacific. 98 Bureau of Rural Sciences

65 Japan's fishery grounds (Takahashi and Yokawa 1999, p. 1). There is a paucity of published information on the activities of the fleet and a logbook program did not commence until Japanese fishers have used large-mesh driftnets to catch tuna and billfish since the early 1970s; the fishery lands swordfish mainly as a bycatch (Sakagawa 1989, pp ). The larger driftnetters operate in international waters and target squid, salmon, pomfret and albacore. A coastal driftnet fishery based in northern Honshu targets striped marlin, but also lands yellowfin, skipjack tuna and swordfish. By the mid-1970s the offshore driftnet fleet was reporting significant catches of swordfish, ranging up to per cent of Japan s total north-western Pacific catch of swordfish. In 1992 the United Nations General Assembly passed a resolution for a moratorium on the use of large-scale (>2.5 km length) pelagic driftnets in international waters (United Nations 1993). In response, Japan restricted driftnet fishing to the waters of its 200 nm exclusive economic zone off Honshu and Hokkaido. Since 1993 Japan s driftnet effort has declined considerably. The number of boats using driftnets in 1994 was 40 per cent of that in 1990, with the catch declining proportionally with fishing effort (Takahashi and Yokawa 1999, p. 2). The Japanese have harpooned swordfish since at least the early 1900s. Striped marlin is the primary target of the harpoon fleet, but swordfish and blue marlin are occasionally taken (Wildman 1997, p. 28). Catches of swordfish by the harpoon fleet ranged from 2557 t in 1952 to 121 t in The decline in catches during the 1970s coincided with increased catches by the driftnet fishery (Sakagawa 1989, p. 69). Catches Catch trends Since 1952 the coastal fleet has landed an estimated t of swordfish per year. 1 The coastal harpoon catch of swordfish had ranged up to 2558 t per year, but then declined after 1970 to t per year. In contrast, coastal driftnet catches increased rapidly in the early 1970s peaking at 3488 t in Driftnet catches then declined, falling to less than 1000 t per year in the 1990s.Coastal longline catches generally increased, from less than 500 t in most years of the 1950s and 1960s to almost 1400 t in In contrast to coastal longline catches, the total number of swordfish landed by offshore and distant-water longliners in the North Pacific increased rapidly during the 1950s, from in 1952 to in During the 1950s fishing effort doubled from 113 million hooks to 220 million hooks per year. After 1963, when longliners changed to day sets for tuna, catches declined to swordfish per year, with effort ranging between 200 and 250 million hooks per year. Catches continued to decline after 1988 in response to further reductions in fishing effort in the North Pacific (<180 million hooks per year) reaching a low of swordfish in The nominal catch rate for swordfish taken by the offshore and distant-water longliners in the North Pacific rose steadily during the 1950s from 1.6 swordfish per 1000 hooks, to peak at 3.6 per 1000 hooks in Longliners targeted swordfish and the fishery s area steadily expanded during the 1950s. Following the change to tuna targeting, catch rates 1 The North Pacific and Pacific-wide swordfish stock assessments use Japan s offshore and distant-water longline data. However, none has incorporated the catches taken by the coastal swordfish fleet, which has included longline, harpoon and driftnet boats. Before 1993, estimates of the annual catch taken by the coastal fleet came from fishery cooperative sales slips. The cooperatives do not record fishing method or area. Through interviews with fishers the National Research Institute of Far Seas Fisheries (NRIFSF) reconstructed catches by gear. Swordfish fisheries 99

66 Driftnet Harpoon Coastal longline Other gear Catch ('000 t) Year 0 Figure 19. Annual catches of swordfish reported by Japan s boats fishing in the north-western Pacific (Dr K. Yokawa, 8 June 1999). declined and fluctuated between 1.0 and 1.8 swordfish per 1000 hooks during the remainder of the 1960s and through the 1970s and 1980s. During the 1990s nominal catch rates declined to less than 1.2 per 1000 hooks. Size composition During swordfish caught by commercial boats in the Higashioki fishing grounds ranged from 60 to 230 cm EFL (~ cm LJFL), with a mode at 140 cm EFL (~160 cm LJFL; Yokawa and Uozumi 1997). Note, however, that commercial fishers tend not to measure small swordfish. Size frequency data from the training and research vessels, which mostly operate outside commercial fishing grounds, show two size modes, with the larger mode at cm EFL (~ cm LJFL) and a smaller one at 60 cm EFL (~70 cm LJFL). Species associations We did not find any published information on the composition of the swordfish directed longline catches. Distribution Prior to 1940 most longliners operated in the waters off the eastern coastline of Japan and some of the larger boats worked the equatorial waters near the Philippines and Indonesia. During Japan s longline fishery expanded across the Pacific, reaching North America s western coastline and extending throughout the tropical and subtropical areas of the North Pacific. During the late 1960s and early 1970s a considerable portion of the fishing effort shifted away from swordfish and the north-western Pacific. Many of Japan s longliners began targeting southern bluefin tuna off southern Australia and New Zealand, striped marlin off Mexico and bigeye in the western and central Pacific. The expansion of the distant-water fisheries was directly related to the development of super-cold (< 50ºC) freezers and targeting of sashimi tuna. In the north-western Pacific, the longliners that target swordfish operate mostly between ºE and 20 45ºN (the Higashioki fishing grounds; Uosaki and Takeuchi 1997). The largest longline catches of swordfish and the highest catch rates occur there. Over the wider northern Pacific, the geographical distribution of high swordfish abundance shifts seasonally (Sosa- Nishizaki and Shimizu 1991). During January it is concentrated between 140ºW and 140ºE (near the western coastline of Japan) and 25 45ºN. After January the area of high swordfish abundance moves away from Japan s coast, occurring east of 150ºW and south to 20ºN during April to June. 100 Bureau of Rural Sciences

67 Japan's fishery The area is smallest during July and August, covering the area from Japan s coastline to 180ºE and 25 45ºN. By September the area expands again, reaching 50ºN and 140 W. Since 1993 new information reported in the logbook has allowed swordfish-targeted effort to be distinguished from tuna-directed effort (Uosaki and Takeuchi 1997). The distribution of swordfish effort varies seasonally, shifting southwards during January March to concentrate in a narrow band at N. It then expanded westwards to 140 E during April June when swordfishtargeted effort peaks. By July September swordfish effort was diffusely spread across the western portion of the fishing ground (24 45 N, E), before slowly contracting into a band at N and spreading eastwards to 180 again (Uosaki and Takeuchi 1997). In contrast to the distant-water longliners, driftnetters operate seasonally. Most driftnet fishing is off Honshu during July October. The harpoon fleet operates in waters off the Bonin and Izu Islands, south of Tokyo and off the northern coastline of Honshu, during December May (Wildman 1997, p. 28). Oceanography The Higashioki fishing grounds coincide with a major frontal zone of the North Pacific, where the warm Kuroshio Current meets the cold Oyashio Current, off the northern part of Japan (around 40ºN). The Kuroshio Current originates from the warm waters of the North Equatorial Current when it turns northwards along the coastline of the Philippines. The Kuroshio flows north-easterly along the coastline of Japan up to 35 40ºN, where it turns eastwards and leaves the coastline to become the Kuroshio Extension (Chikuni 1985, pp. 4 6). Off southern Japan, the Kuroshio extends from the coastline out to about 100 nm. The current flows rapidly (comparable to the Gulf Stream off Florida), although the velocity and volume of water in the current varies seasonally, increasing twice a year (in July September and January March), with longer-term (7 9 year) cycles also observed. Off southern Japan (30 35ºN) the Kuroshio often forms large-scale meanders containing a large cold water eddy at the core surrounded by several small cold or warm water eddies. In July September, the sea surface temperature of the current ranges from 25 to 27ºC. From June to November, a distinct thermocline develops at around 100 m depth. By contrast, a vertical temperature gradient is indistinct and large concentrations of diatoms in the surface layers increase turbidity during January May (Chikuni 1985, pp. 4 6). The Kuroshio Extension maintains the characteristics of the Kuroshio Current to around 170ºE, eventually becoming the North Pacific Current. North of 40ºN, a branch of the Kuroshio Extension mixes with cold Oyashio and subarctic water (Chikuni 1985, pp. 4 6). The Oyashio Current is the southerly flowing convergence of the cold waters of the East Kamchatka Current and the Okhotsk Sea Gyre. The East Kamchatka Current originates in the Bering Sea and flows down the eastern coastline of the Kamchatcka Peninsula. The Okhotsk Sea Gyre is a western branch of the East Kamchatka Current that is further cooled in a counterclockwise revolution of the Okhotsk Sea. The Oyashio Current has a permanent, sharp gradient in salinity or halocline at a depth of m. In the upper layer, a marked thermocline develops during July September. Cooling and vertical mixing occurs during January March but generally stops at the depth of the halocline. The sea surface temperature of the current ranges from 6 10ºC in July September to 1 2ºC in January March (Chikuni 1985, pp. 4 6). In the convergence zone off northern Japan (30 45ºN, ºE) the waters of the Oyashio Current sink under the Kuroshio and they rapidly mix and form numerous cold and warm eddies. The formation of eddies is highly variable and some are occasionally huge, moving away from the boundary of the currents to the north (warm Kuroshio origin) or south (cold Oyashio origin; Chikuni 1985, pp. 4 6). Gear and targeting Fleets Three boat types comprise Japan s longline fleet: Swordfish fisheries 101

68 coastal longliners (10 20 GRT), which specifically target swordfish in coastal waters of the Higashioki fishing grounds; larger ( GRT) offshore longliners, which operate in offshore waters of the Higashioki grounds and in tropical areas of the western Pacific; and distant-water longliners ( GRT), which operate throughout the world s oceans, fishing for extended periods by superfreezing catches at below 50ºC. The distant-water longliners are highly mobile, fishing in different areas at different times of the year to take advantage of the best fishing conditions for the various species targeted. In 1980 there were 1171 distant-water longliners and about 1000 offshore longliners. The fleet had declined to 743 distant-water and 371 offshore longliners by 1995 (Wildman 1997, p. 49; Miyabe and Bayliff 1987, p. 7). In contrast to the larger longliners, the coastal fleet appears to have remained fairly stable in numbers, totalling 821 boats in 1980, falling to 536 in 1986 and then recovering to 819 boats in 1994 (Wildman 1997, p. 52). During the late 1990s a fleet of offshore longliners targeted mainly swordfish. Most are based in the port of Kesennuma. Fishing trips are usually 40 days in duration and involve about 25 longline sets. They move seasonally, following the southern side of the frontal zone between the Kuroshio and Oyashio current. The swordfish longliners sometimes target blue shark or tuna, such as bigeye, yellowfin or northern bluefin, during October February (Dr K. Yokawa, 2 March 2000). Boats that are mostly larger than ten GRT comprise the fleet of billfish driftnetters. Fleet size peaked in 1989 at 468 boats and then declined steadily in response to the United Nations ban on driftnets. In 1994 only 123 boats remained in the fleet. The harpoon fleet consists of around 100 boats that range in size from 3 to 15 GRT (Wildman 1997, p. 28). Fish-finding We did not locate any published information on techniques used by Japan s longliners to find swordfish in the North Pacific. Our review of Australia s fishery describes fish-finding, fishing gear and practices used by Japan s longliners operating off eastern Australia (p. 138). Fishing gear and practices To target swordfish, offshore longliners set their gear shallow at night, deploying 3 4 branchlines between buoys. During fishing trips they frequently switch from night sets for swordfish (or blue shark) to day sets, using about 12 hooks per basket for tuna. Squid is the preferred bait for swordfish. However, the longliners sometimes use mackerel, which is cheaper and often produces diverse catches of tuna, shark and swordfish (Dr K. Yokawa, 2 March 2000). The branchlines are also much shorter than those used to target bigeye or yellowfin. The deepest hooks of swordfish sets may reach a depth of 170 m compared with 300 m or more for sets targeting bigeye. The offshore and distant-water longliners deploy about 2000 hooks per set over a distance of km. 102 Bureau of Rural Sciences

69 Japan's fishery Japan s longliners use traditional rope and basket gear. Nylon mainlines were introduced to the directed swordfish fishery around Japan in the early 1990s (Takahashi and Yokawa 1999, p. 2). In 1995, however, logbook reports indicated that most swordfish operations used traditional rope gear (only 18% of the swordfish directed sets used nylon mainlines and 34% used nylon branchlines; Uosaki and Takeuchi 1997). Most now use rope gear because of difficulties in handling monofilament gear and reports that it does not improve catch rates (Dr K. Yokawa, 1 March 2000). The Japanese rarely use lightsticks (Miyabe 1996, p. 1). Driftnet boats deploy monofilament nets with a mesh size of 18 cm. The driftnets are about 9 m deep and up to 12 km long. Typically, the nets are set in the afternoon and retrieved before dawn (Wildman 1997, p. 28). Harpooners sight swordfish from the crow s nest and throw the harpoon by hand. The harpoon is tipped with an electric dart that delivers a fatal shock to the fish (Wildman 1997, p. 28). The offshore longliners that target swordfish store their catches on ice and land it fresh. It is sold on local, niche sashimi markets. The Japanese process swordfish in several ways, depending on the size of the fish. Coastal and offshore longliners usually remove the head, viscera and fins of small (<30 kg W wt ) swordfish, whereas they fillet larger swordfish. Some boats remove the bill (not the head), viscera and fins. Distant-water longliners land mostly fillets, depending on the availability of fish hold space. The range and variation in processing methods create complexities for estimating catch weights. Discarding of target species We did not locate any published information on the bycatch or discards of Japan s longline fishery in the North Pacific. Anecdotal reports suggest that longliners land about 95 per cent of the swordfish that they catch (Dr K. Yokawa, 1 March 2000). Off eastern Australia, Japan s longliners retained most (94%) of the swordfish caught. In the western Pacific, large numbers of blue, mako and thresher shark are caught by longliners of several fleets. The Japanese do not report shark to species level in their national logbooks (Bailey et al. 1996, p. 4.13). Assessment Data collection The National Research Institute of Far Seas Fisheries is responsible for collecting and processing data from Japan s offshore and distant-water fleets. Based on interviews with fishers and fishery cooperative sales reports, the Ministry of Agriculture, Fisheries and Forestry staff estimate catch levels of the domestic swordfish fleets, including the coastal longline, driftnet and harpoon fleets (Miyabe 1996, p. 2). The institute commenced a logbook program for the coastal swordfish fishery in Since 1952 they have collected logbooks from offshore and distant-water longliners (Uozumi and Yokawa 1997, p. 2). The National Research Institute of Far Seas Fisheries added details of fishing gear and operations, including the number of hooks between buoys, to the logbook in It revised the logbook s format in 1993, incorporating information on processed weights and the type of materials used for the mainline and branchlines. The current logbook collects information on the boat, crew and daily activities, including date, noon position, sea surface temperature, number of baskets, number of hooks and the catch (number and weight of each species retained; Miyabe 1996, p. 1). Fishers must declare whether the majority of sets on a trip targeted swordfish, tuna or shark. In the late 1990s the coverage of the logbooks was around 50 per cent for the coastal boats, 80 per cent for offshore longliners and 95 per cent for the distant-water longliners (Uozumi and Yokawa 1997, p. 2; Takahashi and Yokawa 1999, p. 2). Prior to 1994 Japan estimated the catch weight of swordfish from a small number of length frequency samples provided by commercial fishers (who tend not to measure smaller swordfish). The size data had a 10º by 20º resolution. Since 1994 Japan has used processed weights reported in logbooks. However, Japanese fishers land swordfish in several states; some swordfish are gilled and gutted, whereas others are filleted (see p. 102). Processed weight is converted to whole weight using a conversion curve applied to size composition data by area. The conversion curve is Swordfish fisheries 103

70 determined by research cruises that roughly accounts for the various processed states of swordfish (Dr. K. Yokawa, 24 March 2000). Details of the catch and effort of the large-mesh driftnetters have been catalogued in compulsory logbooks since The information recorded is much the same as for the longliners, with details of the gear including the number of TAN (length unit of the driftnet, usually 30 m), length of TAN, mesh size and details of the catch in number and weight by species (Miyabe 1996, p. 2). Japan has monitored the size composition of swordfish catches since Longliner landings are sampled by three methods: government sampling at major unloading ports; wholesaler sales slips (detailing dressed weights only); and on-board measurements by commercial, training and research vessels. There are, however, several important limitations to the size data set. Historically, the collection of swordfish size data in the Pacific Ocean has been very poor, with coverage amounting to less than 0.2 per cent of the total commercial catch during the late 1980s and 1990s and less in earlier years (Yokawa and Uozumi 1997). Coverage of the research and training vessel catches has been little better, at less than 1.5 per cent of the total catch. Sampling size by sex has only occurred since 1986, with about 70 per cent of the size data recorded with information on sex. Commercial fishers typically only measure the processed weight of fish (Miyabe 1996, p. 1) and they tend to not measure the small fish (Yokawa and Uozumi 1997). Most of the size data has come from the training and research vessels, which almost always operate in areas outside the commercial fishing grounds. There has been no observer validation of any of the commercially derived swordfish catch or size data. Japan weighed driftnet-caught swordfish at the major unloading ports from No size sampling of the driftnet fleet catch has occurred since 1995, because the effort and distribution of the fleet has declined. The catch of the harpoon fleet has not been sampled. Research Swordfish are often alive when landed. Japan has a long history of biological research on billfish, including swordfish. Fishery research stations and maritime colleges of most prefectures in Japan have training vessels operating commercial longline fishing gear. The training vessels routinely record the species and size composition of their longline catches and stomach contents and sex of individual fish (Yokawa and Uozumi 1997, p. 1). Several early analyses (e.g., Koga et al.1972) describe the distribution of fishing grounds in relation to environmental conditions and billfish biology (size composition, feeding, reproduction and assumed migration patterns). Nishikawa et al. (1985, pp ) summarise surveys of fish larvae, which have recorded swordfish larvae at various locations. More recent fieldwork (e.g., Nakano et al.1997, p. 43) has investigated the effects of water temperature and longline configuration on catch composition and catch rates. Japan is also undertaking a population genetics study of swordfish (Uozumi and Yokawa 1997, pp. 3 4). Stock assessment Fishery indicators In 1989 Skillman (1989, p. 188) estimated the average weights of swordfish on a Pacific-wide basis. Annual average weights showed a sudden sharp increase between 1961 and 1962, which is when Japan s longliners began setting their longlines around dawn, leaving the longline to soak during the day. Those day sets seemed to catch fewer, but larger swordfish. After 1962 the trend was towards smaller sizes, although large fluctuations occurred. Skillman interpreted the decrease in the 1960s and 1970s as a possible expression of increased fishing pressure. However, the calculation of average weights was based on the number of swordfish reported in logbooks; that 104 Bureau of Rural Sciences

71 Japan's fishery was raised to total numbers for the Pacific (to account for other fleets), using Food and Agriculture Organization estimates of Pacific landings (weight). As swordfish sizes appear to vary with latitude, the calculation of annual average weights on a Pacific-wide basis, using data where the distribution of effort varied considerably over latitudes during the time series, may well produce results that do not reflect real trends. Bartoo and Coan (1989) made a qualitative assessment of the status of Pacific swordfish by examining nominal catch rate trends in conjunction with a stratified methodology, which allowed the separation of areas with high catch rates (assumed to represent swordfish target areas) from areas with low catch rates. They examined a single, Pacific-wide and a three-unit (north, south and east) stock hypothesis, using Japan s offshore and distant-water longline data for On a Pacific-wide stock basis, the catch rate in swordfish target areas increased slowly through the 1950s and then abruptly dropped by about one-third in 1963 when tuna operations commenced. It remained static until In areas where swordfish were mostly caught as a bycatch of the tuna operations, the catch rate trend was fairly stable, increasing only marginally over the time series. Those areas did not appear to have been affected by the change in fishing gear and operations seen in the swordfish target areas. When examined as a three-stock case, the trends in north-western stock catch rate patterns followed the Pacific-wide case because the bulk of the catches and hence data, came from the area. Bartoo and Coan concluded that, until 1980, the stocks had not been exploited heavily enough to cause a noticeable decline in catch rates. More recently, Nakano (1998) attempted to infer the status of Pacific swordfish stocks from trends in standardised catch rate of Japan s offshore and distant-water longline fleet during Using a generalised linear model, he standardised for gear, season, area and target species effects. As new data regarding the gear configuration were available since 1975, the analyses were carried out for two sets of data; and Japan s longliners have set hooks progressively deeper since 1974 and, in the North Pacific, Nakano found that gear effects were negatively correlated with hook depth; that is, catch rate decreased as hook depth increased (Nakano 1998, p. 8). The standardised catch rates in the North Pacific peaked in By 1968 it had declined to less than half the 1952 rate. Catch rates gradually increased again but, until 1983, catch rate remained below the levels observed in the early 1950s. Since then it has fluctuated, peaking in 1986 and then declining and fluctuating around levels observed during the 1970s. The trends observed in the North Pacific were considerably different from those in the south-western and eastern stock unit an observation supporting the multi-stock hypothesis. Swordfish fisheries 105

72 New logbooks, introduced in 1993, have allowed the standardisation of catch rate to incorporate the effects of different fishing gear and operations, particularly the number of hooks between buoys, which directly affects hook depth. While the time series of the data was too short to elucidate any trends ( ), the variance in catch rate caused by gear effects was large, highlighting the importance of collecting and incorporating data on the effects of different fishing gear and operations (Uosaki and Takeuchi 1997). Generalised linear models have also been used to standardise catch rate data for the large-mesh driftnet fishery in offshore waters of the North Pacific. Although fluctuations occurred, Uosaki (1997) suggests that no clear trend was observable. Stock assessment models Previous stock assessments of the Pacific swordfish resource have utilised Japan s offshore and distant-water longline fishery data, because it is the only data from the Pacific that is suitable in both detail and time period for catch rate analyses. Japan s offshore and distant-water longline data include the catch in numbers of fish and the nominal fishing effort (hooks), by five-degree square, for each month since Only production models have been used because the lack of comparable and reliable size frequency data has precluded the use of age-structured models for assessments. Sakagawa and Bell (1980) assessed Pacific swordfish stocks by applying an equilibrium production model to longline data for a single, Pacific-wide and a three-stock hypothesis. They were able to generate a solution for data only, which corresponds to the period when the fishery targeted swordfish with night sets and only for the Pacific-wide stock hypothesis. They estimated the maximum sustainable yield to be t per year with an optimum effort of 2.2 million hooks per five degree square. The average yield between 1966 and 1975 was around t, produced by 1.8 million hooks per five degree square, which was considerably below the levels estimated to produce maximum sustainable yield. The fishery was probably more efficient at catching swordfish in those earlier years, before longliners targeted tuna with day sets. They concluded that the stock appeared healthy and capable of sustaining increased yields with increased effort. A precautionary note was given suggesting that, if the fishery recommenced night swordfish operations, the increased efficiency of the gear could result in catches exceeding the maximum sustainable yield if the level of fishing effort increased by only 25 per cent. Skillman (1989) made a similar assessment of the theoretical Pacific-wide swordfish stock from data, using a generalised equilibrium production model where the population dynamics were assumed to be density dependent. When no change in gear efficiency was allowed for, he estimated a maximum sustainable yield of about t (range t) with an optimal effort of around 200 million hooks. He also presented another scenario, where the change to deep longlining in 1974 was factored to have caused a 25 per cent reduction in catch rate for swordfish. The incorporation of the effects of deep longlining increased the estimates of maximum sustainable yield to t. Skillman concluded that the fishery was operating near the point of maximum sustainable yield at the beginning of the 1980s. Recently, two stock assessments have examined a hypothetical North Pacific stock. Yokawa (1999) and Nakano (1998) both used two data sets from Japan s longline fishery; and The data contained additional information on the configuration of longline gear, including the number of hooks between buoys, which was used to distinguish target species. The catch rates were standardised by sub-areas of similar average catch rates and trends. Yokawa used smaller sub-areas than those used by Nakano. Yokawa estimated an aggregated annual catch rate series as the sum of sub-area standardised catch rates, weighted by the relative area of each sub-area to the total stock area. His approach resulted in low catch rates during the 1950s when very high catch rates were recorded in very small areas. The high catch rates in small areas may have distorted the catch rate trends in Nakano s standardisation towards greater values. Yokawa s analyses revealed an essentially flat long-term trend and the estimates of recent catch rate were comparable to 40 years ago. Yokawa (1999) used his standardised catch rates in an ASPIC non-equilibrium production model. The results were largely indeterminate because the model required one or more parameters to be 106 Bureau of Rural Sciences

73 Japan's fishery fixed to achieve a solution. However, results indicated that swordfish biomass has been at levels above maximum sustainable yield (Yokawa 1999, Figures 20 22). The difficulties in fitting the model may be caused by a lack of contrast in the catch and effort data, e.g., low exploitation rates for North Pacific swordfish where the change in catch rate as a result of fishing effort is so small that no relationship to catch can be observed. However, problems in fitting the model can also arise from inaccurate total catch data, incorrect assumptions about the stock boundaries or a lack of adjustment for environmental, gear (such as bait) or other significant effects in the catch rate standardisation. Kleiber (1999) applied two slightly different production models to Japan s and Hawaii s longline data. The hypothetical stock boundaries were north of 15 N and west of 135 W. One of the models used was a modified Pella Tomlinson model, with extensions that allowed difference in the two fleets, variable catchability, carrying capacity and a cryptic catch by undocumented fisheries. The second model allowed the separation of net production factors into recruitment and natural mortality. Both models were largely inconclusive, predicting a range of contrasting scenarios, varying from declining to rising abundance trends, slow to fast turnover rates and minimal to very high exploitation rates. As reported for Yokawa s assessment, there were difficulties in fitting the model and achieving convergence on parameter estimates, perhaps due to a lack of contrast in the data. Status The stock structure of swordfish throughout the Pacific is unclear, although the possibility of separate north and south stocks and potentially south-eastern and south-western stocks, exists. Despite this uncertainty, assessments of the different stock hypotheses have reached similar conclusions: that the stocks in the Pacific were underexploited until the early 1980s. Assessments of data beyond 1980 have been inconclusive and the condition of the swordfish stock remains uncertain. Recent standardisation of catch rates in Japan s fishery has revealed that the catch rate has remained stable since the mid-1970s and has been linearly related to the total catch (Yokawa 1999). Assuming that catch rate in Japan s fishery reflects swordfish abundance, this suggests that the North Pacific swordfish is currently exploited at a low rate (Wildman 1997, p. 10) and that abundance has been high natural variability. An alternative conclusion is that the standardised catch rates are not an accurate indicator of abundance and that the status of the stock is uncertain (Dr J. Joseph, 7 May 2000). Reliability of the assessment The early assessments did not include catches of the coastal fleets and consequently may have underestimated maximum sustainable yield. Since the early 1950s coastal fleet catches have ranged from 1200 to 4000 t per year or per cent of Japan s North Pacific swordfish catch. In 1999 the Swordfish Working Group of the Interim Scientific Committee for Tuna and Tunalike Species in the North Pacific Ocean (ISC) reviewed recent attempts at swordfish stock assessment. It identified the following stock assessment-related areas of research: Development of production models that include small-scale spatial differences in stock dynamics would lead to more realistic assessments than present models that assume stock homogeneity. More accurate models of the relationship between catch rate and abundance need to be developed. Time and area of fishing, gear and practice changes, fleet dynamics and the influence of oceanographic and environmental phenomena, need to be better accounted for so they can be more accurately factored into catch rate standardisation models. The effects of age-structure and sex-structure on production models need to be determined for swordfish, which has a complex pattern of life-history dynamics due to the sexual dimorphism of the species (ISC 1999). The collection of catch data needs to improve for many of the nations fishing swordfish in the North Pacific if future assessments are to be robust. Furthermore, a lack of size composition data for the fishery has precluded age-structured assessments. Many years of reliable size data are Swordfish fisheries 107

74 needed to produce a time series suitable for a robust assessment and a comprehensive and long term size sampling program should be implemented immediately. Management Institutions The Fisheries Agency of Japan (FAJ) manages Japan s tuna and billfish fisheries. Japan has no domestic regulation for the swordfish fishery operating in the Pacific Ocean. Japan is an active member of the Inter-American Tropical Tuna Commission, but there are currently no region-wide regulations that apply to swordfish in the Pacific Ocean. Coastal and fishing nations, including Japan, are negotiating the creation of a regional fishery management arrangement covering western and central Pacific tuna fisheries (the Multilateral High Level Consultation or MHLC). The commission eventually created by MHLC might share responsibility with Inter-American Tropical Tuna Commission for the North Pacific swordfish fishery. Bycatch Bailey et al. (1996) review information on bycatch in western and central Pacific tuna fisheries, including Japan s longline fisheries. Observers have been placed on Japan s offshore and distantwater longliners calling at ports in Australia, New Zealand, the Federated States of Micronesia and Kiribati (Bailey et al. 1996, p. 4.5). Those longliners mostly targeted bigeye and yellowfin. Therefore the observer information on bycatch is not directly relevant to Japan s North Pacific swordfish fishery. Anecdotal reports indicate that offshore longliners that target swordfish with night sets take a wide variety of bycatch species, mostly albacore, bigeye, yellowfin, striped marlin, shark (blue, salmon, mako), opah, pomfret, mahi mahi, lancetfish and gemfish. The longliners land almost all of those species expect for lancetfish and gemfish (Dr K. Yokawa, 2 March 2000). Interaction Recreational anglers do not target swordfish in Japan. But, there is potential for interaction between Japan s harpoon, driftnet and longline fleets targeting swordfish in coastal waters of the Higashioki fishing grounds. Indeed, declines in swordfish by the harpoon fleet during the 1970s corresponded to increased catches by the driftnet fishery. It is unclear, however, whether the decline is evidence of the harpoon and driftnet fisheries directly competing for the local stock of swordfish (Sakagawa 1989, p. 69) or whether it reflects a changeover from harpoon to driftnet fishing gear. In the wider North Pacific there is potential for interaction between the various nations taking swordfish. Several nations target swordfish (e.g., Hawaii-based longliners and Taiwan s coastal and offshore fleets) whereas others have a substantial bycatch of swordfish, e.g., Taiwan s distantwater longliners. Advice needs There have been several analyses of the population structure of Pacific swordfish using mtdna variation. However, there is still no conclusive evidence of population structure. A basin-wide tagging program, in conjunction with analysis of genetic and meristic variations should be implemented to conclusively delineate stock structure. Such a program would also provide useful estimates of fishing mortality rates, natural mortality rates and growth rates. The effects of large-scale oceanographic phenomena like the El Niño Southern Oscillation events and decade scale changes in ocean productivity on swordfish populations, fisheries and assessments should be further examined. Future prospects Japan s market provides stable domestic demand for frozen swordfish. Demand and prices paid for frozen swordfish are increasing gradually, but they are lower than prices for frozen sashimi tuna, such as yellowfin. While swordfish prices are not competitive with tuna, it is unlikely the major distant-water longline fleets will redirect effort at swordfish. There are concerns, however, that were swordfish prices to rise significantly or catch levels to decline in other fisheries, 108 Bureau of Rural Sciences

75 Japan's fishery longliners from Taiwan, Korea or the European Union might move to the Higashioki fishing grounds and compete with Japan s longliners for swordfish. The fresh-chilled swordfish landed by offshore longliners is eaten as sashimi in a few areas, such as Kesennuma. The prices paid for fresh swordfish at those markets is comparable to the prices paid in the United States. Consequently, exporting swordfish to the United States is not commercially attractive at present (Dr K. Yokawa, 2 March 2000). Swordfish is reported to have excellent qualities for use as sashimi. However, the populations of larger cities of Japan generally do not recognise swordfish as a sashimi species, resulting in lower demand (and prices) outside of niche markets. In recent years, sashimi markets have suddenly developed for albacore, which were previously not considered a sashimi species. Consequently, there is potential for a promotion of sashimi swordfish to suddenly and massively expand the domestic market for swordfish (Dr K. Yokawa, 2 March 2000). References Bailey, K., Williams, P.G. and Itano, D. (1996) By-catch and discards in western Pacific tuna fisheries: A review of SPC data holdings and literature. South Pacific Commission, Nouméa. Oceanic Fisheries Program Technical Report No. 34. Bartoo, N.W. and Coan, A.L. Jr (1989) An assessment of the Pacific swordfish resource. In Stroud, R.H. (ed.) Planning the Future of billfishes: Research and management in the 90s and beyond. Proceedings of the Second International Billfish Symposium, Kailua Kona, Hawaii, August National Coalition for Marine Conservation, Savannah. Chikuni, S. (1985) The fish resources of the northwest Pacific. FAO Fisheries Technical Paper 266. Food and Agriculture Organization of the United Nations, Rome. ISC (1999) Report of the Swordfish Working Group Meeting, January 1999, Honolulu, Hawaii. Document ISC2/99/PLEN/11submitted to the Second Meeting of the Interim Scientific Committee for Tuna and Tuna-like Species in the North Pacific Ocean (ISC). Honolulu, Hawaii, January Kleiber, P. (1999) Very preliminary North Pacific swordfish assessment. Document ISC2/99/2.2 presented at the Second Meeting of the Interim Scientific Committee for Tuna and Tuna-Like Species in the North Pacific Ocean (ISC). Honolulu, Hawaii, January Koga, S., Imanashi, H. and Tawara, S. (1972) The fishing conditions of the tuna and marlin in the central South Pacific. Journal of the Shimonoseki University of Fisheries 20(3), pp Not seen, abstracted in Green, R. (1988) Bibliography of Japanese literature on the tuna (excluding southern bluefin tuna) and billfishes of the Coral and Tasman Seas. CSIRO Marine Laboratories Report No Miyabe, N. (1996) Information paper on the data collecting system for the Japanese swordfish fisheries. In Hinton, M.G. (ed.) Working Group for the Collection of Statistical and Biological Information on Pacific Swordfish: Report of the First Meeting, July Scripps Institution of Oceanography, La Jolla, California. Miyabe, N. and Bayliff, W.H. (1987) A review of the Japanese longline fishery for tunas and billfishes in the eastern Pacific Ocean, Inter-American Tropical Tuna Commission Bulletin 19(1), pp Nakano, H. (1998) Stock status of swordfish in the Pacific Ocean inferred from standardised CPUE of the Japanese longline fishery using general linear models. In Barrett, I, Sosa Nishizaki, O. and Bartoo, N. (eds) Biology and fisheries of swordfish, Xiphias gladius: Papers from the International Symposium on Pacific Swordfish, Ensenada, Mexico, December United States Department of Commerce, NOAA Technical Report NMFS 142. Swordfish fisheries 109

76 Nakano, H., Okazaki, M. and Okamoto, H. (1997) Analysis of catch depth by species for tuna longline fishery based on catch by branch lines. Bulletin of the National Research Institute of Far Seas Fisheries 34, pp Nishikawa, Y., Honma, M., Ueyanagi, S and Kikawa, S. (1985) Average Distribution of Larvae of Oceanic Species of Scombrid Fishes, Far Seas Fisheries Research Laboratory, Shimizu. S Series 12. Sakagawa, G.T. (1989) Trends in fisheries for swordfish in the Pacific Ocean. In Stroud, R.H. (ed.) Planning the Future of billfishes: Research and management in the 90s and beyond. Proceedings of the Second International Billfish Symposium, Kailua Kona, Hawaii, August National Coalition for Marine Conservation, Savannah. Sakagawa, G.T. and Bell, R.R. (1980) Swordfish, Xiphias gladius. In Shomura, R.S. (ed.) Summary Report of the Billfish Stock Assessment: Pacific Resources. United States Department of Commerce, NOAA Technical Memorandum NMFS SWC5. Skillman, R.A. (1989) Pacific billfish stocks. In Stroud, R.H. (ed.) Planning the Future of billfishes: Research and management in the 90s and beyond. Proceedings of the Second International Billfish Symposium, Kailua Kona, Hawaii, August National Coalition for Marine Conservation, Savannah. Sosa Nishizaki, O. and Shimizu, M. (1991) Spatial and temporal CPUE trends and stock unit inferred from them for the Pacific swordfish caught by the Japanese tuna longline fishery. Bulletin of the National Research Institute of Far Seas Fisheries 28, pp Takahashi, M. and Yokawa, K. (1999) Brief description of Japanese swordfish fisheries and statistics in the Pacific Ocean. Document ISC2/99/SWO1 presented at the meeting of the Swordfish Working Group of the Interim Scientific Committee for Tuna and Tuna like Species in the North Pacific Ocean, January United Nations (1993) United Nations General Assembly Resolution 46/215, effective 1st January Uosaki, K. (1997) The CPUE trend for the North Pacific swordfish caught by the Japanese largemesh driftnet fishery. Paper presented at the Second International Pacific Swordfish. Symposium. Kahuku, Hawaii. March Uosaki, K. and Takeuchi, Y. (1997) CPUE Standardisation of Pacific Swordfish Using the Japanese Longline Data Based on New Format Logbook. Paper presented at the Second International Pacific Swordfish Symposium, March , Kahuku, Hawaii. Uozumi, Y. and Yokawa, K. (1997) Country Report of Japan for the Swordfish Fisheries in the Pacific Ocean. Working document presented at the Second International Pacific Swordfish Symposium, March , Kahuku, Hawaii. Wildman, M.R. (1997) World Swordfish Fisheries: An analysis of swordfish fisheries, market trends and trade patterns, past-present-future, Volume III Asia. United States Department of Commerce, NOAA Technical Memorandum NMFS F/ SPO 25. Yokawa, K. (1999) Standardised catch rate for swordfish caught by Japanese longliner in the North Pacific and the tentative trial of production model. Document ISC2/99/SFWTG/2.1 presented at the Second Meeting of the Interim Scientific Committee for Tuna and Tuna-like Species in the North Pacific Ocean (ISC). Honolulu, Hawaii, January Yokawa, K. and Uozumi, Y. (1997) Evaluation and Preliminary Analysis of Size Data of Swordfish Caught by the Japanese Longliners During Working document presented at the Second International Pacific Swordfish Symposium, March , Kahuku, Hawaii. 110 Bureau of Rural Sciences

77 Hawaii Hawaii s swordfish fishery Overview Hawaii-based longliners target swordfish or tuna (predominantly bigeye). The swordfish longliners catch mostly large (61 77 kg) swordfish in cool (13 19 C) waters of the subtropical and subarctic frontal zones of the North Pacific Ocean. The fishery features long fishing trips (>30 days), with longliners regularly venturing nm from Hawaii. The number of longliners targeting swordfish increased rapidly in , then declined. The swordfish catch peaked at almost 6000 t in The catch then declined to less than 3200 t per year. Better economic returns for tuna longlining were largely responsible for the decline in swordfishdirected longlining and the subsequent fall in swordfish catch levels. Many longliners consider that targeting swordfish is more risky because it involves longer trips and higher costs than tuna longlining. Public concern over incidental catches of shark and protected seabirds and turtle is the main force currently driving management. The fishery Development Immigrant Japanese started longlining for tuna in Hawaii in 1917 using techniques adopted from Japan. Those earliest activities used high-bowed, wooden Sampan boats from the local pole-andline fishery for skipjack tuna (Boggs and Ito 1993, p. 70). The fishery developed steadily, supplying tuna to local sashimi markets and canneries, then declined during (Dr C. Boggs, 23 September 1999). Hawaii-based longliners had an incidental catch of swordfish, ranging up to 17 t in 1987 (Ito and Coan 1999, p. 7), but then they did not commence targeting swordfish until In that year, exploratory longlining along banks off the Northwestern Hawaiian Islands successfully adapted techniques from the Florida-based swordfish fishery (Ito et al. 1994, p. 3). Those techniques involved shallow, night sets with squid baits and lightsticks. Fishers immediately realised that large quantities of swordfish could be caught, stored on ice and profitably airfreighted to mainland United States. Swordfish longlining was an instant success and by the end of 1989 at least ten longliners were fishing for swordfish for at least part of the year (Ito et al. 1994, p. 12; Dollar Year 0 Figure 20. Annual number of Hawaii-based longliners targeting swordfish (Ito et al. 1994, p. 40; Ito and Machado 1997, p. 20; Ito and Coan 1999, p. 13; Mr P. Dalzell, 21 July 1999). Swordfish fisheries 111

78 1991, p. 1). By Hawaii-based longliners were fishing for swordfish (Ito et al. 1994, p. 12; Figure 20). The influx of longliners to Hawaii during the early 1990s was akin to the Californian gold rush of 1849 (Travis 1999, p. 93). Most of the longliners that entered the fishery were from mainland United States. Those from the United States east coast relocated to Hawaii because of increased competition and declining economic returns in the Gulf of Mexico shrimp fishery and the Gulf swordfish fishery (Townsend and Pooley 1994, p. 297; Ito et al. 1994, p. 3).Other longliners moved from the east coast to Hawaii because of increasingly stringent regulations in the Atlantic swordfish fishery (Travis 1999, p. 90). Of the east coast boats, 24 were relatively large, sophisticated longliners (Townsend and Pooley 1994, p. 297) that we term tuna boats. Former shrimp trawlers and longliners ( Gulf boats ) arrived from the Gulf of Mexico during At first the Gulf boats targeted yellowfin and bigeye. They then began targeting swordfish in 1992 (Ito et al. 1994, p. 3; Townsend and Pooley 1994, p. 297). About 20 longliners from the United States west coast also relocated to Hawaii in (Travis 1999, p. 91). Many of the Gulf boats returned to the Gulf of Mexico in There are various explanations given for their departure from Hawaii. Vietnamese-American crew members had strong family ties and many wished to be closer to their families in the United States mainland. Some Gulf boats wanted to ensure a stake in the Atlantic swordfish fishery that was about to be closed to new entrants. Others were attracted to California, where operating costs were marginally lower than Hawaii. Thirteen Gulf boats temporarily relocated to Fiji. Certainly profits for Gulf boats and other longliners targeting swordfish fell sharply in Nine of the thirteen returned to Hawaii or mainland United States in and the remaining four were sold to pay off debts or confiscated (Travis 1999, pp. 94, 97). Many tuna boats continued to target bigeye during the development of the swordfish fishery. In the early 1990s some of the tuna boats that had targeted swordfish switched back to targeting tuna, which provided more reliable economic returns. By the mid-1990s few tuna boats were targeting swordfish, leaving that component of the fishery to the Gulf boats. Each year fewer than 50 longliners have targeted swordfish since In 1998 low prices for swordfish resulted in more longliners changing their operations to target tuna. Other swordfish longliners moved to California earlier in July September and stayed there longer because of savings in airfreight and fuel costs. Catches Catch trends The Hawaii-based fishery expanded rapidly, from a catch of 281 t of swordfish for 1989, to 1901 t for 1990, peaking at 5942 t for 1993 (Ito and Coan 1999, p. 7; Figure 21) when swordfish longliners reported setting 3.8 million hooks (Mr R. Ito, 15 July 1999). 1 Annual catches then declined to 2726 t by 1995 (Ito and Coan 1999, p. 7) and remained under 3200 t per year for the remainder of the 1990s (Mr R. Ito, 15 July 1999). Size composition Medium and large-sized swordfish characterise the Hawaii-based fishery. During landed swordfish averaged 70 kg, spanning a wide size range, from less than 25 kg to more than 475 kg (Ito and Machado 1997, p. 42). Before swordfish targeting had become 1 All weights are whole or round weights derived by multiplying trunk weights by 1.45 (Ito, 18 May 1999). Published catch statistics might not include unknown quantities of small or damaged swordfish that are discarded at sea or swordfish caught by Hawaii-based longliners but landed on the United States west coast. 112 Bureau of Rural Sciences

79 Hawaii important in the late 1980s the average weight of swordfish landed by Hawaii-based longliners was kg. In 1994, at the height of the swordfish fishery, landed swordfish averaged 78 kg. In later years the average weight fluctuated between 71 and 78 kg. Size frequencies presented by Ito and Machado (1997, p. 42) show that small (<25 kg) swordfish dominated longliner landings prior to the increase in swordfish targeting in the late 1980s, e.g., they accounted for more than 30 per cent of the 331 swordfish recorded in Since then small swordfish are rare or absent in landings data. The relative scarcity of small swordfish in recent data might be due to longliners discarding them at sea or to the effects of a 50 nm closure around the Northwestern Hawaiian Islands or it might simply reflect the dominance of tunadirected operations in the data Year 0 Figure 21. Annual landings of swordfish by Hawaii-based longliners (WPRFMC 1998, p. 3-58; Ito and Coan 1999, p. 7; Mr P. Dalzell, 21 July 1999). Species associations From observer data, NMFS (1998) estimated that Hawaii-based longliners caught more than shark in Incidental catch of shark, particularly blue shark, is more common in longline sets targeting swordfish than in those targeting tuna (Ito and Machado 1997, p. 7; McCoy and Ishihara 1999, p. 5). Longliners targeting swordfish catch sharks (50% of the total catch by number), swordfish (28%) and albacore (11%). Other species include mahi mahi (4%), bigeye (3%), marlin (2%) and yellowfin (1%; Ito and Machado 1997, p. 36). Distribution Following the fishery s development and the exclusion of longlining near the Hawaiian Islands in 1991 (see p. 121), the main centre of longline activity shifted northward and westward of Hawaii s exclusive economic zone (Ito et al. 1994, p. 39). Average number of miles to first set, a measure of how far the longliners range, increased from 585 nm in 1991 to 884 nm in 1995 (WPRFMC 1998, p. 3-22). By 1998 most of the swordfish longlining was in international waters, north of Hawaii s zone to 40 N and between 140ºW and 180 W. Seasonal weather conditions affect the seasonal and geographical distribution of longliner activity. Several Hawaii-based longliners are capable of fishing year-round. Most, however, range widely during July September when seas are calm, longlining closer to the Hawaiian Islands at other times (Ito et al. 1994, p. 10). Many of the longliners that target swordfish also move seasonally to base operations in ports on the United States west coast, such as Long Beach and San Pedro Swordfish fisheries 113

80 during July September. The five returns of swordfish tagged suggest a cyclical pattern to the movements of adult swordfish in the North Pacific; some moved east to be recovered off California days after release, while others were recaptured about one year near the site of release in the central North Pacific. No tagged swordfish was reported recaptured to the west, near Japan (Dr C. Boggs, 23 September 1999). Fishers report a lunar pattern to their swordfish catches: catch rates are highest a few days before and after full moons (Bigelow et al p ). Large, female swordfish are most common during full moon periods (Mr B. McNamarra, 24 July 1999). Oceanography For Japan s swordfish fishery we describe the oceanography of the North Pacific Ocean with particular emphasis on the north-west Pacific(Map 6). Waters of the central-north Pacific where Hawaii-based longliners target swordfish can be divided into three major oceanographic regions: the subarctic domain which extends northward of about 43ºN, characterised by a cool (10 14ºC), low salinity surface layer that is about 120 m deep in January March and 30 m in July September; the subtropical domain which extends southward from 31ºN to about 20ºN, characterised by a warm (~25ºC), saline surface layer that is about 125 m deep in January March and 60 m in July September; and the North Pacific Transition Zone which occupies the region between the subarctic and subtropical domains. Temperature and salinity gradually increase southwards in the North Pacific Transition Zone (Roden 1991, pp. 2 3; Laurs and Lynne 1991, pp ). Two frontal zones delineate the oceanographic regions described above: -50 Kuroshio Current Kuroshio Extension North Pacific Current Hawaii North Equatorial Current E 170W Map 6. The North Pacific Ocean, showing the location of major ocean circulation systems (Roden 1991, p. 2). White indicates ocean depths of less than 1000 m, light grey indicates depths of m and dark grey indicates depths greater than 2000 m. characterised by strong sea surface temperature gradients in January March, the Subarctic Frontal Zone ranges between 40ºN to 43ºN; and 114 Bureau of Rural Sciences

81 Hawaii the Subtropical Frontal Zone occurs between about 31ºN and 34ºN as a temperature front during November May and as a salinity front throughout the year. Sea surface temperature gradients of 3ºC/50 nm characterise the Subtropical Frontal Zone, with steeper gradients occurring across convoluted eddies associated with the zone (Roden 1991, pp. 3 4). The distribution of the swordfish fishery is closely linked to the position of the North Pacific Transition Zone and catch rates depend on the strength of the associated fronts. During July September, catch rates are highest at the Subarctic Frontal Zone when sea surface temperatures rise to 13 19ºC. Longline fishers believe that swordfish follow squid, such as the flying squid, as they migrate from the Subtropical Zone to the Subarctic Zone in July September, returning to the Subtropical Zone in October December (Dr C. Boggs, 23 September 1999). Swordfish catch rates tend to be highest when sea surface temperatures of the Subtropical Frontal Zone fall to 16 19ºC during November May (Bigelow et al p. 193). Gear and targeting Fleets Two types of Hawaii-based longliner have targeted swordfish: former shrimp trawlers and yellowfin longliners ( Gulf boats ) from the Gulf of Mexico that are primarily operated by Americans of Vietnamese descent; and purpose-built longliners operated by Americans of Korean descent and Caucasian Americans from Hawaii, the west coast of the United States and the North Atlantic seaboard ( tuna boats ). Together, tuna boats and Gulf boats are referred to as swordfish longliners when they are targeting swordfish. 2 There are also a few small, local tuna longliners, which opportunistically target swordfish. Most of the longliners fishing for swordfish are m long and all are steel-hulled. Long trips of more than 30 days duration are a feature of the Hawaii-based fishery (Ito et al. 1994, pp. 7, 12). The average number of days of fishing per trip has increased steadily, from 10.7 in 1991 to more than 13.0 in 1995 (WPRFMC 1998, p. 3-22). In 1996 longliners targeting swordfish averaged 7.7 trips per year with 18 days travel and 14 days fishing per trip (Folsom et al. 1997, p. 67). They thus average 246 sea days per year. Fish-finding When targeting swordfish, Hawaii-based longliners seek sea surface temperatures of C. They search for waters with strong sea surface temperature gradients and current shear. Local bathymetry is important near the Hawaiian Islands, with longliners concentrating on known rises (deep sea canyons and seamounts). Wider out, in the northern convergence zones, longliners operate over a fairly featureless ocean floor where the water is more than 4000 m deep (Bigelow et al. 1999, p. 22). The tuna boats use radar, Loran and satellite or global positioning systems (GPS) for navigation and temperature probes, automated track plotting and satellite weather imaging to help them find swordfish associated with thermal fronts (Ito et al. 1994, p. 8). In contrast, the Gulf boats, which are now the main fleet targeting swordfish, have a low-technology approach to longlining. They monitor sea surface temperature, but few regularly access satellite imagery for information on sea surface temperature, ocean colour or ocean production (Mr B. McNamara, 23 July 1999). Both types of longliner use radio beacons, strobe lights and radar reflectors to locate their fishing gear (Boggs and Ito 1993, p. 73). 2 Longlining for swordfish is a sub-component of the Hawaii-based longline fishery. Some longliners target swordfish while other longliners target tuna, such as bigeye and yellowfin. A few sometimes target both swordfish and tuna during the same trip. Analyses of catch rate data differentiate trip types according to gear and operational data gathered from skippers or crew members at dockside interviews. Shallow, night sets using lightsticks and short, weighted branchlines characterise swordfish trips. Tuna trips use line-shooters to position the moderate to deep, day sets with no lightsticks (He et al. 1997). Mixed trips deploy sets of both types. Swordfish fisheries 115

82 Fishing gear Before 1985 Hawaii-based longliners used Japan-style rope gear to target swordfish. They introduced monofilament longline gear and line setters in 1985 and, by 1991, almost all Hawaiibased longliners used monofilament gear (Townsend and Pooley 1994, p. 297). The monofilament mainline is stored on a hydraulic-powered reel (Ito et al. 1994, p. 7). Their branchlines are usually m long (Table 8). They attach g lead weights to branchlines, about 5 7 m above the hook and lightsticks are attached 2 3 m above the hook (Mr B. McNamara, 23 July 1999). Fishing practices Longliners targeting swordfish usually commence setting their gear at, or just before, sunset. Sometimes they set their gear after sunset when seabirds are less active in the area (Mr J. Cook, 12 February 1999). The number of hooks per set has not changed significantly since swordfish targeting commenced in Longliners targeting swordfish averaged hooks per set during (Ito and Machado 1997, p. 22). To increase catch rates of swordfish, they use a rubber band to fasten a lightstick to branchlines about two metres above hooks (Ito et al. 1994, p. 9). Between sets, fishers vary the frequency of lightsticks depending on catch rates and catch composition. Longliners targeting swordfish generally use two types of gear configuration derived from the United States North Atlantic fishery. The Florida or east coast style has two or three branchlines between buoys, with a lightstick attached to each branchline (Bigelow et al. 1999). Gulf boats set 3 5 branchlines between each buoy, with one lightstick almost every second branchline (Mr B. McNamarra, 24 July 1999). Unlike tuna longliners, the longliners that target swordfish do not use mainline shooters. The mainline is set close to the sea surface to target swordfish; during setting the mainline is freespooled with slight tension on the hydraulic reel maintaining tension on the line (Ito et al. 1994, p. 8). Branchlines are usually 20 m long and buoy lines are 9 18 m. Fishers believe that hooks usually lie m below the sea surface (Folsom et al. 1997). In 1998 longliners started using baits dyed a deep blue colour. This is believed to reduce the incidental catch of seabirds the bait is more difficult for the birds to see and it may increase catch rates (Ito et al. 1994, p. 9). The increased catch rates might be due to dyed baits being somehow more attractive to swordfish. Alternatively, the increased catches might be an artifact of reduced bait loss to seabirds. Longline setting takes 4 6 hours. After setting, the longline is left to soak for 6 10 hours before hauling (Bigelow et al. 1997, p. 6). Hauling usually begins before sunrise. Duration of the haul varies depending on the number of fish caught and the sea conditions. Most swordfish are dead when retrieved. Discarding of target species Hawaii-based longliners sometimes encounter patches of small swordfish (<23 kg). Small swordfish and badly damaged swordfish have little or no commercial value and are usually consumed by crew members, given away when the boat returns to port or, more rarely, discarded at sea (Ito et al. 1994, p. 16). 116 Bureau of Rural Sciences

83 Hawaii Marine mammal (killer, false killer and pilot whale) sometimes remove or damage fish that are hooked on the longline. In contrast, damage by blue shark and mako shark is very common. Longliners will often move large distances away from areas where shark damage is high in relation to swordfish catch rates. Cookie cutter shark often bite into hooked swordfish and tuna. Dead swordfish usually float to the sea surface in the North Pacific. 4 Bird (and squid) may then enlarge shark wounds on hooked swordfish floating at the surface. Assessment Data collection The United States National Marine Fisheries Service (NMFS) Honolulu Laboratory is responsible for monitoring the Hawaii-based longline fleets. Fishery statistics are available from four sources: longline logbooks; market sampling; State of Hawaii catch reports; and observer monitoring (Ito and Coan 1999, p.5). Since 1987 NMFS in conjunction with the Hawaii Division of Aquatic Resources has conducted twice-weekly market sampling that provides data on the number of longliners, their fishing trips and the processed weight and price of each fish landed (Ito et al. 1994, Table 8. Summary of gear characteristics of Hawaii-based longliners that target swordfish (Dollar 1991; 3 Mr B. McNamara, 23 July 1999). Type Characteristics Boat size m Type of gear monofilament stored on hydraulic reels Length of mainline km Diameter of mainline mm Distance between buoys m Length of branchline ~13 m Diameter of branchline ~2.1 mm Type of hooks 8/0 10/0 Mustad Number of hooks per set (range), 724 (mean) Branchlines per buoy 3 5 Type of lightsticks Cyalume (break activated) and World Plastics (thaw activated) Type of bait used shortfin squid pp. 5 6; Ito and Coan 1999, p. 3; Ito and Machado 1997, pp. 3 5). Sex-specific data is not available because most swordfish are landed as processed trunks (with the head, gills, viscera and fins removed). Weights are raised to whole weights (Ito and Coan 1999, pp. 3 5). Based on dockside interviews with skippers or crew members, NMFS staff allocate a trip type ( swordfish, tuna or mixed ) to each trip. When dockside interviews are not available, they determine trip types from operational data reported in logbooks (Ito and Machado 1997, p. 4). Logbooks have been mandatory since November 1990 (Ito and Coan 1999, p. 5). For each day, fishers record the location, times of longline set and retrieval, gear configuration, number of hooks set, number of each species caught, oceanographic conditions and interactions with protected species (Ito and Coan 1999, p. 5; Ito and Machado 1997, p. 3). 3 Cited in Ito et al. (1994, p. 29). 4 The reason dead swordfish float in some areas and not in other areas is unclear. It might relate to differences in water temperature and salinity that influence water density and thus buoyancy. Swordfish fisheries 117

84 Landings are usually estimated from the number of swordfish retained (logbook data) multiplied by the average weight (landings data; Ito et al. 1994, p. 6). The landing estimates do not include fish discarded at sea (Mr B. McNamarra, 24 July 1999). An observer program began in November 1990 when longliners volunteered to take NMFS observers to investigate interactions with Hawaiian monk seals (Ito and Coan 1999, p. 5). In 1994 NMFS established a mandatory observer program to collect information on interactions with protected turtle, discarding and fishing operations (Ito and Machado 1997, p. 1). The program was later extended to seabird interactions and observers are now placed on about 5 per cent of longliner trips (Ito and Coan 1999, p. 5). In there was a research program, involving observers, to educate fishers on the environmental impacts of their activities (Ito and Machado 1997, p. 2) and to develop bycatch mitigation measures (Mr B. McNamarra, 24 July 1999). Research Since 1992 the United States Congress has appropriated funds each year to the Pelagic Fisheries Research Program (PFRP) to undertake research activities related to the conservation and management of pelagic fish species for the Western Pacific Regional Fishery Management Council (WPRFMC). The program is multidisciplinary, involving short to medium-term projects based in various research organisations both in Hawaii and internationally. It has supported a variety of fishery data collection programs and biological, oceanographic and socioeconomic research and stock assessment on pelagic fisheries of United States western Pacific territories and states. Many of those projects are focused on Hawaii s fisheries and several are relevant to the swordfish fishery, e.g., fishing boat economics, identification of oceanographic features critical to pelagic fish distribution, genetic analysis and archival tagging of swordfish (PFRP 1999). Since 1991 the NOAA research vessel Townsend Cromwell has dedicated one or two research cruises each year to collecting data on swordfish biology and ecology. The cruises target swordfish with standard monofilament longline gear. They use hooks timers and time depth recorders to collect data on swordfish behaviour in relation to the depth configuration of longlines. Staff monitor oceanographic conditions and take biological measurements and samples from catches. Samples include muscle tissue (for genetic analyses), gonads (reproduction), stomachs (diet) and hard parts (age and growth). Live swordfish are often tagged and released. Commercial longliners sometimes also tag and release swordfish under a voluntary tagging program (Ito and Coan 1999, pp. 4 5). Stock assessment Fishery indicators Nominal catch rates reported by longliners targeting swordfish have shown no clear trend since the fishery began in the late 1980s. Catch rates rose from an average of 9.0 swordfish per 1000 hooks in 1987 to 15.4 in 1991 (Figure 22). They then fell to 10.3 per 1000 hooks in 1994, but increased in subsequent years. In 1997 swordfish catch rates returned to the 1991 level (15.4 swordfish per 1000 hooks), falling slightly to 14.5 in 1998 (Ito and Coan 1999, p. 3). 118 Bureau of Rural Sciences

85 Hawaii 20 Swordfish trips Mixed trips Year Figure 22. Nominal catch rates of swordfish reported by Hawaii-based longliners (WPRFMC 1998, p. 3-61; Mr P. Dalzell, 21 July 1999). Although nominal catch rates have shown no clear trend, longline fishers acknowledge that they are continually accumulating experience in locating and catching swordfish. Some fishers believe that nominal catch rates may further mask declining abundance of swordfish because longliners are ranging greater distances and searching for longer periods. The additional searching effort is not incorporated in the single measure of effort (total hooks) used in calculating catch rates. Bigelow et al. (1997) 5 used generalised additive models to adjust catch rates for various environmental factors and variations in longlining practices. Those factors accounted for more than 40 per cent of the variation in swordfish catch rates. Standardised catch rates, for example, increased with latitude until N and increased near temperature fronts and during full moon periods. The standardised catch rates declined from 1991 to late 1994, recovering slightly during Dissipation of the Subtropical Frontal Zone in 1994 may have contributed to those low catch rates (Bigelow et al. 1999, pp. 24, 46). Bigelow et al. (1999) did not adjust catch rates for improvements in experience and efficiency that probably occurred over the study period. Increases in experience and efficiency would accentuate the declines in adjusted catch rates. Improved experience and knowledge may have led longliners to fish in the subtropical and subarctic frontal zones. Thus, time and area variables used in the analysis would represent some of the improvements in experience. The average whole weight of swordfish landed by swordfish longliners ranged between 70 and 81 kg during (Figure 23). The fleet-wide average was smaller and varied between 54 kg in 1987 and 81 kg in 1992 (Ito and Machado 1997, p. 33). However, care needs to be taken in drawing conclusions from those size data because they do not take into account the unknown levels of discarding of small swordfish. 4 5 Bigelow et al. (1999) use swordfish sets identified by He et al. (1997) through cluster analysis of species composition data. Swordfish fisheries 119

86 Average whole weight (kg) Year Figure 23. Average weights of swordfish landed by Hawaii-based longliners that targeted swordfish (WPRFMC 1998, p. 3-17; Mr P. Dalzell, 21 July 1999). Stock assessment models Kleiber s (1999) analysis of catch and effort data aggregated by five degree square month from Hawaii-based longliners and Japan s longliners for the entire North Pacific is described on page 107. NMFS is currently developing an operational model to evaluate the performance of various stock assessment methods and management options against uncertainties in data, biological parameters and fleet responses. The underlying stock assessment models incorporate attributes of agestructured and length-based models that take growth, reproduction, mortality, recruitment and exploitation into account (Dr M. Labelle, 11 May 1999). Status Kleiber s assessment was inconclusive. Both models produced a range of contradictory scenarios from declining to rising abundance trends, from slow to fast turnover rates and from minimal to very high exploitation rates (Kleiber 1999, p. 6). Bigelow et al. s (1997) adjusted catch rates declined until late 1994, but those analyses covered a short period of the fishery ( ). Some fishers believe that fishery indicators, such as nominal catch rates and average weights, mask a decline in the abundance of swordfish in the North Pacific. They do not accept the explanation of a broad-scale change in oceanographic conditions influencing swordfish availability and thus catch rates. Reliability of the assessment The models of Bigelow et al. (1999) provide insights on fishery performance, but do not necessarily track changes in swordfish abundance. Their models could be applied to the longer time-series of Japan s longline data for the North Pacific to track changes in fishery performance. Kleiber partly attributed the poor fit of his models to a lack of contrast in the data, which is consistent with a low exploitation rate of swordfish. Critical to his assessment are assumptions about stock structure and mixing rates. Tag recapture studies would help to distinguish the stock s boundaries and provide information on mixing rates, which could be used in spatially disaggregated models Bureau of Rural Sciences

87 Hawaii Management Institutions The Western Pacific Fishery Management Council sets management policies for Hawaii s fisheries that are then approved by the National Marine Fisheries Service. Legislation requires wide public consultation for the development of management policy. Amendments to the council s Pelagic Fisheries Management Plan are virtually all due to longline fishery amendments 2 5 and 7. Objectives Under the United States Magnuson Stevens Act, all fisheries must be managed within maximum sustainable yield (or a proxy to maximum sustainable yield). In 1987 the National Marine Fisheries Service implemented a fishery management plan for pelagic fisheries in the 200 nautical mile exclusive economic zones of United States possessions, states and territories of the western and central Pacific. The plan s objectives are to: manage pelagic fisheries to achieve optimum yield; promote domestic harvest and values; reduce gear conflict; improve the statistical base for stock assessment and fishery evaluation; promote regional and international arrangements for assessing and conserving the stocks; preclude waste associated with fishing; and promote domestic marketing in United States territories and possessions (WPRFMC 1998, p. 1). Surplus production models Also known as biomass dynamic models, surplus production models are a mathematical representation of the way a stock of fish responds to the removal of individuals, e.g., by fishing. The models are based on the theory that, at large stock size, reproductive rates and rate of stock growth are slowed by self-regulating mechanisms (e.g., predation) and that stock growth rates are faster after removals, because the stock attempts to rebuild. In theory, fishing can be moderated to take advantage of the more productive stock growth rates, provided that it does not exceed the stock recovery capacity. Production models assume that there is one equilibrium population size, where gains from recruitment and growth balance losses from mortality, to which the population will naturally return. They assume that catch rates are an index of abundance. Those models also require a wide range of fishing effort (maximum sustainable yield cannot be precisely estimated if effort has not exceeded the optimum level). The latest models, such as ASPIC, allow some restraints to be relaxed (e.g., equilibrium population size) and take into account changes in the population s age structure or the distribution of fishing activity. Tuna longliners switching to swordfish targeting and the activation of latent fishing effort are potential problems to the management of Hawaii s swordfish fishery. Other management issues include interactions with seabirds and turtle, gear conflict between various users and the finning of shark (Ito and Machado 1997, p. 1 3). Management measures Since 1990 all longliners have been required to have a federal longline fishing permit which includes information on boat dimensions and characteristics (Ito et al. 1994, p. 7). In 1991 federal regulations prohibited longlining within 50 nm of the Northwestern Hawaiian Islands to prevent interactions with the endangered Hawaiian monk seal. Federal regulations prohibiting longline fishing within 50 to 75 nm of the main Hawaiian Islands were also introduced in 1991 to prevent gear conflict with smaller trolling and handline boats (Folsom et al. 1997, p. 72). The arrival of the Gulf boats in Honolulu produced considerable friction with local fishers (Travis 1999, p. 91). In 1991 to arrest the rapid increase in longliner numbers, the Western Pacific Fishery Management Council limited the total number of longliners to 166 (Townsend and Pooley 1994, p. 299). The limit applies to all Hawaii-based longliners, many of which targeted swordfish. Yet, longline activity continued to expand after the council closed the fishery, with swordfish effort peaking in Townsend and Pooley (1994, p. 299) suggest that limiting the fishery deferred the making decisions on important management issues, such as the optimum level of effort for the fishery and management responses to excess effort. A moratorium allowed one permit transfer during and restricted replacement to new boats having no more fishing Swordfish fisheries 121

88 power than the replaced boat. Few operators relinquished their permits or replaced existing boats with new longliners. Townsend and Pooley (1994, pp ) conclude that the restrictions severely limited business decisions and froze further development and rationalisation of the longliner fleet. Using a spatial-dynamic model, Chakravorty et al. (1999, p. 85) suggest that area closures are more effective in controlling catch levels of target species, such as swordfish, than is revenue taxation. In 1994 the Western Pacific Fishery Management Council replaced the moratorium on new entrants with a three-year limited entry program that capped the fishery at 164 longliners (Ito and Coan 1999, p. 3) and allowed unlimited transfer of permits (Dr C. Boggs, 23 September 1999). In 1998, 110 of those permits were actively used. Of those, about 30 were used by longliners targeting swordfish. In 1996 NMFS instigated a pilot program that equipped 112 longliners with a satellite-based vessel monitoring system that facilitates the monitoring of fishing activities in closed areas. Enforcement agencies used the system to identify several infringements. More importantly, vessel monitoring systems are considered a major deterrent to longlining in closed areas, greatly reducing the costs of surveillance and enforcement (Ito and Machado 1997, pp. 2 3). United States-flagged boats operating in international waters are subject to the provisions of the United States High Seas Fisheries Compliance Act, which requires that all boats hold a permit and complete a federal logbook or an acceptable equivalent, such as the NMFS longline logbook. Since 1991 the Western Pacific Fishery Management Council has enforced regulations and placed observers on United States boats fishing outside the exclusive economic zone. However, recently it has become apparent that the terms of the Magnuson Stevens Act might only allow the council to regulate activity within the zone (Mr P. Dalzell, 21 July 1999). Bycatch Swordfish longlining has a greater interaction with turtle and seabird than does tuna longlining, which involves day sets in tropical waters. Turtle and seabird are a concern in Hawaii, particularly in the swordfish fishery. Seabirds with high rates of fishery interaction are the and the Laysan albatross. Based on observer data, Kleiber (1999) 6 estimated that Hawaii-based longliners hooked about 1963 black-footed albatross and about 1479 Laysan albatross in Of those, nearly 80 per cent were released dead (NMFS 1998). Longliners occasionally encounter other seabird, such as shearwater, booby (near landmasses) and fulmer. The turtle species of major concern is the leatherback turtle. Only small numbers are taken, but concern arises from the low population of the species. Longliners also incidentally take olive Ridley, loggerhead and green sea turtle (in order of greater number taken). Lawsuits filed against the National Marine Fisheries Service in late 1999 resulted in a federal district court judge ruling that the agency must update the Environmental Impact Statement (EIS) for the Hawaii-based longline fishery, including the impacts of longlining on turtle and other protected species. The judge ordered the temporary closure of a large area ( ºW, north of 28ºN) of Hawaii's longline fishery to reduce longliner interaction with leatherback turtle. Subsequently, the judge proposed 100 per cent observer coverage of longlining activities and a greater reduction in the number of longline sets. At the time of writing the judge had stayed the order for two weeks pending meetings by the parties. A new lawsuit filed in July 2000 seeks to halt fishing by longliners based in California because of interactions with protected species and recreational angling (Dr M. Laurs, 11 July 2000). Blue shark are the most numerous species caught by longliners targeting swordfish. Crew members used to remove the fins of shark and discard the carcass. In 1999 about 67 per cent of shark caught by Hawaii-based longliners were finned, but they retained few of the carcasses (Dr M. Laurs, 11 July 2000). Five Hawaii-based longliners had permits to receive and land shark fins from foreign longliners outside Hawaii s exclusive economic zone (Ito and Machado 1997, p. 3). In 2000 the State of Hawaii banned the landing of shark fins (Dr M. Laurs, 11 July 2000). 6 Cited in Garcia and Associates (1999, p. i) 122 Bureau of Rural Sciences

89 Hawaii Interaction Occasionally seabirds have been washed up on beaches after swallowing lightsticks. Lightsticks are believed to present a more significant danger to seabird chicks flying fish eggs, which are a food source for several seabird species, attach to floating debris such as lightsticks. The seabirds may accidentally swallow the lightsticks and regurgitate plastic fragments when feeding their chicks, which may choke the chick. Surveys of Laysan albatross in the Northwestern Hawaiian Islands found that nearly 98 per cent of the albatross had ingested more than 23 grams of plastic each. Lightsticks were the source of some, but not all, of that plastic (WPRFMC 1999). Advice needs Marine wildlife concerns appear to be more important than stock status in managing the fishery. One fishery manager suggested, for example, that the death of a single short-tail albatross could close the fishery. Bycatch of turtle is a concern in several longline fisheries. There is a general belief amongst fishery managers and scientists that the swordfish stock is in a healthy state. Current catch levels are half the historical peak catch and fishery indicators (e.g., standardised catch rates and size composition) do not show any clear trend that would suggest a stock-wide problem with swordfish in the North Pacific. The decline in Hawaiibased swordfish catches is attributed to social and economic causes. Nevertheless, a reliable stock assessment is needed to verify that the status of North Pacific swordfish is indeed healthy. Such an assessment would need to take into account the cumulative effects of past catches, the combined effect of catches in other fisheries (e.g., California) on a common stock and the major changes in the geographical distribution of longline activities. A tag recapture experiment would be useful in providing a point estimate of population size and population parameters, such as natural mortality, fishing mortality and growth. It would also provide useful information on stock structure and mixing, which are critical sources of uncertainty in Kleiber s (1999) assessment. However, there are practical difficulties in tagging the numbers of swordfish that such an experiment requires. Large swordfish caught on longlines are valuable and live swordfish are aggressive and difficult to tag. More importantly, they are usually dead or moribund, even when retrieved within a few hours of the longline set. Future prospects The swordfish subcomponent is considered a stable fishery. Some participants believe that the Hawaii-based swordfish fishery will stabilise at about t per year. Such speculation, however, ignores market prices and swordfish availability as the factors determining catch levels. As a result of a restaurant boycott and increasing supplies of swordfish from South America and Australia, swordfish prices declined during Consequently, more longliners switched to targeting tuna and swordfish trips are unlikely to resume until prices improve. References Bigelow, K.A., Boggs, C.H. and He, X. (1999) Environmental effects on swordfish and blue shark catch rates in the US North Pacific longline fishery. Fisheries Oceanography 8(3), pp Swordfish fisheries 123

90 Boggs, C.H. and Ito, R.Y. (1993) Hawaii s pelagic fisheries. Marine Fisheries Review 55(2), pp Chakravorty, U., Nemoto, K. Kinping, T. and Yanagida, J. (1999) A spatial-dynamic model for the allocation of fishing effort: Application to the Hawaii longline pelagic fishery. pp In Chakravorty, U. and Sibert, J. (eds) Ocean-scale management of pelagic fisheries: Economic and regulatory issues. Proceedings of an international workshop organized by the Pelagic Fisheries Research Program, November SOEST JIMAR Contribution Joint Institute of Marine and Atmospheric Research, University of Hawaii at Manoa, Honolulu, Hawaii. Dollar, R.A. (1991) Summary of Swordfish Longline Observations in Hawaii, July 1990 March Southwest Fisheries Science Centre Administrative Report H National Marine Fisheries Service, Honolulu. Folsom, W.B., Crory, D.M. and Brewster-Geisz, K. (1997) North America Swordfish Fishing. in World swordfish fishing: An analysis of swordfish fishing operations, past-present-future (Volume IV). United States Department of Commerce, NOAA Technical Memo NMFS F/SPO 28. Garcia and Associates (1999) Hawaii Longline Seabird Mortality Mitigation Project. Draft Final Report to Western Pacific Regional Fishery Management Council, Honolulu, Hawaii. He, X., Bigelow, K.A. and Boggs, C.H. (1997) Cluster analysis of longline sets and fishing strategies within the Hawaii-based fishery. Fisheries Research 31, pp Ito, R.Y. and Coan, A.L. Jr (1999) U.S. swordfish fisheries of the North Pacific Ocean. Working Document presented at the Second Interim Scientific Meeting of the Committee for Tuna and Tuna-like Species in the North Pacific Ocean (ISC). Honolulu, Hawaii. January Ito, R.Y., Dollar, R.A. and Kawamoto, K.E. (1994) The Hawaiian longline fishery for swordfish. Proceedings of the International Symposium on Pacific Swordfish: Development of Fisheries, Markets and Biological Research. Ensenada, Baja California, Mexico, December Ito, R.Y. and Machado, W.A. (1997) Annual report of the Hawaii-based longline fishery for Southwest Fisheries Science Center Administrative Report H National Marine Fisheries Service Honolulu Laboratory Honolulu, Hawaii. 124 Bureau of Rural Sciences

91 Hawaii Kleiber, P. (1999) Very Preliminary North Pacific Swordfish Assessment. Working document submitted at the Second Meeting of the Interim Scientific Committee for Tuna and Tuna-like Species in the North Pacific Ocean (ISC). Honolulu, Hawaii, January Laurs, R.M. and Lynn, R.J. (1991) North Pacific albacore ecology and oceanography. pp In Wetherall, J.A. (ed.) Biology, oceanography and fisheries of the North Pacific Transition Zone and Subarctic Frontal Zone. Papers from the North Pacific Transition Zone Workshop. Honolulu, Hawaii, 9 11 May NOAA Technical Report NMFS 105. United States Department of Commerce. McCoy, M.A. and Ishihara, H. (1999). The socioeconomic importance of sharks in the U.S. flag areas of the western and central Pacific. Southwest Fisheries Science Centre Administrative Report AR SWR National Marine Fisheries Service, Honolulu. PFRP (1999) Pelagic Fisheries Research Program < (accessed on 10 August 1999). Roden, G.I. (1991) Subarctic Subtropical Transition Zone of the North Pacific. Large-scale aspects and mesoscale structure. pp In Wetherall, J.A. (ed.) Biology, oceanography and fisheries of the North Pacific Transition Zone and Subarctic Frontal Zone. Papers from the North Pacific Transition Zone Workshop. Honolulu, Hawaii, 9 11 May NOAA Technical Report NMFS 105. United States Department of Commerce. Seki, M.P., Polovina, J.J., Kobyayashi, D.R. and Mundy, B.C. (1999) The oceanography of the Subtropical Frontal Zone in the central North Pacific and its relevancy to the Hawaii based swordfish fishery. Working document ISC2/99/3.3 presented at the Second Interim Scientific Meeting of the Committee for Tuna and Tuna-like Species in the North Pacific Ocean (ISC). Honolulu, Hawaii, January Travis, M.D. (1999) Entry and exit in Hawaii s longline fishery, : A preliminary view of explanatory factors. pp In Chakravorty, U. and Sibert, J. (eds) Ocean-scale management of pelagic fisheries: Economic and regulatory issues. Proceedings of an international workshop organized by the Pelagic Fisheries Research Program, November SOEST JIMAR Contribution Joint Institute of Marine and Atmospheric Research, University of Hawaii at Manoa, Honolulu, Hawaii. Townsend, R.E. and Pooley, S.G. (1994) Effort reduction under limited entry; attrition versus fractional licences in the Hawaii longline fleet. North American Journal of Fisheries Management 14, pp WPRFMC (1998) Pelagic fisheries of the western Pacific region 1997 Annual Report. Draft 10/15/98. Western Pacific Regional Fishery Management Council, Honolulu, Hawaii. WPRFMC (1999) Pacific Islands Fishery News, Winter Western Pacific Regional Fishery Management Council, Honolulu, Hawaii. WPRFMC (2000) Council finalizes measures to manage sharks; agrees to measures protecting NWHI fisheries, coral reefs, indigenous fishing rights. Press Release, 2 March Western Pacific Regional Fishery Management Council, Honolulu, Hawaii. Swordfish fisheries 125

92 Australia s swordfish fishery Overview For more than 40 years, Japan s longliners have caught swordfish as a bycatch of fishing operations targeting tuna in what is now the eastern Australian fishing zone (AFZ). Swordfish are also a bycatch of Australia s longliners which target premium quality bigeye, yellowfin and southern bluefin tuna in cooler waters off the south-eastern coastline (35 40 S). They are rare in longline activities targeting bigeye and yellowfin off north-eastern Australia (10 20 S). In many of Australia s longliners relocated from New South Wales to south-eastern Queensland where they used night-set, squid baits to target swordfish and bigeye. Landings of swordfish increased to 2373 t in Basking swordfish are very rare in Australia s eastern waters because of the weak thermocline. Australia s longliners catch swordfish in waters with relatively warm (21 29 C) sea surface temperatures. However, the swordfish mostly live in deeper water that is colder than the surface water. There is a close association with bigeye and the fishery is considered a swordfish bigeye fishery. There are three geographical components to Australia s swordfish fishery: inshore activities along the boundaries of oceanographic fronts and eddies generated by the warm East Australian Current; inshore activities associated with seamounts; and offshore activities associated with oceanic fronts and seamounts. The Fishery Development Expansion of Japan s longline fishery from coastal activities in Japan to distant-water operations was essentially a post-war development. Japan s longliners first started fishing off eastern Australia in 1953 (Anonymous 1953, p. 3). By 1960 they had conducted commercial operations over most of the tropical and subtropical areas that are still fished today. Longlining in Australia s waters quickly became an important component of Japan s global fisheries for southern bluefin tuna and tropical tuna. Globally, the Japanese targeted albacore for canning until the mid-1960s. As the supply of albacore from Taiwan s and Korea s longliners began to increase, Japan s longliners modernised their operations and took advantage of high prices paid for good-quality tuna at sashimi markets in Japan (Caton and Ward 1996, p. 10). In the early 1970s activities spread southwards for premium-quality southern bluefin tuna in temperate oceans (40 50 S). At the same time, Japan s longliners began targeting bigeye and yellowfin in tropical waters, including the Coral Sea. Longlining in what are now the eastern waters of the 200 nm AFZ 1 was a small component of Japan s activities over a much wider region. In the 1970s markets in Japan also developed for billfishes, particularly striped marlin (Caton and Ward 1996, p. 10). Swordfish had commercial markets in Japan, but it was not particularly valuable and Japan s longliners rarely targeted it. During Australians experimented with small-boat longlining in what is now the eastern AFZ. The longline fishery expanded rapidly after 1984 following the successful airfreighting of fresh-chilled yellowfin and bigeye to the sashimi markets of Japan (Caton et al. 1998, p. 18). By 1987, 133 of Australia s longliners were active, reporting a total catch of 998 t D wt of yellowfin. Activity then declined when many longliners stopped fishing because of variable yellowfin catch rates, high freight charges and unpredictable prices paid for the tuna. Since the mid-1990s activities have extended to the occasional targeting of southern bluefin tuna off south-eastern New South Wales (NSW) from about June to September. Swordfish, however, were a minor, incidental catch of Australia s longliners (annual catches ranged up to about 70 t per year). 1 We define the eastern AFZ as Australian fishing zone waters east of 140ºE and north of 40ºS, including AFZ waters around Lord Howe and Norfolk islands. 126 Bureau of Rural Sciences

93 Australia South Equatorial Current AFZ Coral Sea Cairns Fiji -20 Queensland Mooloolaba Brisbane NSW East Australian Lord Howe Island New Caledonia Lord Howe Rise Norfolk Island -30 Tasmania Current Tasman Sea New Zealand -40 West Wind D rift Map 7. The eastern AFZ showing the location of major ocean circulation systems (Ward et al. 2000, p. 101). White indicates ocean depths of less than 1000 m, light grey indicates depths of m and dark grey indicates depths greater than 2000 m. The early longline activities concentrated within 100 nm of the New South Wales coastline (29 39 S). In the early 1990s a second region for Australia s longliners developed in the north-west Coral Sea near Cairns (16 S, 148 E), where there are high catch rates of bigeye and yellowfin. Swordfish are rarely targeted in the area. Since the early 1980s Australian scientists had noted that catch rates reported by Japan s longliners indicated a high abundance of swordfish in the eastern AFZ. During Williams (1993) conducted trials with a commercial longliner using lightsticks. In 25 longline sets, swordfish catch rates on hooks with lightsticks (6.66/1000 hooks) were about twice those for hooks without lightsticks. The swordfish catch rates were substantially higher than swordfish catch rates reported by Japan s longliners (about one swordfish/1000 hooks). Williams project did not, however, stimulate the development of a swordfish fishery at the time. Swordfish fisheries 127