2. THE FISHERIES AND THEIR CATCHES

2.1 North Pacific

The North Pacific region supports some of the most widely publicised of the large scale driftnet fisheries. These fisheries are targeted principally on Pacific salmon (Oncorhynchus spp.), the neon flying squid (Ommastrephes bartrami), and on various tunas and tuna-like species (Scombroidei). The major nations involved in these fisheries are Japan, Korea and Taiwan (Province of China).

The history of the development of large scale driftnet fisheries in the North Pacific is obscure. It is clear, for example, that driftnets were being used in both China and Japan many centuries ago for the capture of a variety of fish species. Quite when these may have moved far enough offshore, or have employed enough netting, to be considered ‘large scale’ or ‘pelagic’ fisheries, is impossible to say. Suzuki (1990), for example, states that the large meshed driftnet fishery for tunas, which operates now in the mid North Pacific, dates back approximately a hundred years, but that is was confined to coastal waters until the mid 1970's. The pelagic driftnet fishery for salmon evidently dates from earlier decades. It has been stated (Anon 1989), that as early as the 1930's, fishermen in Alaska noted the presence of Japanese high seas driftnetters off Bristol Bay, pursuing salmon. The squid driftnet fishery is undoubtedly the most recent of the three major driftnet fisheries, having been initiated in the late 1970's. These three fisheries are considered in turn below.

2.1.1 The salmon driftnet fishery

There are, in fact, several driftnet fisheries for salmon in the North Pacific, but by far the largest of these are Japanese. Coastal driftnet fisheries are operated by US and Canadian vessels in coastal waters and estuaries from Washington State to Alaska. These fisheries are controlled by numerous effort and gear regulations, with a few thousand vessels participating, although most are restricted to the use of less than around 500 m of netting (O'Hara et al 1986, Douglas 1989).

The major driftnet fisheries are Japanese, and there are two of these. A land-based fishery operates in the waters to the Northwest of Japan (Figure 1), and a mothership fishery operates further from Japanese waters, and is now restricted to two areas in international waters of the North Pacific and Bering Sea (Figure 2). The mothership fishery is operated by catcher vessels (and scout vessels), which are supported by larger factory ships (motherships), which process the catch and support the catcher vessels. Each mothership supports around 40 catcher vessels.

Fishing effort in the mothership fishery expanded after the Second World War, which lead, in 1952, to its regulation by the International North Pacific Fishery Commission and by a bilateral USSR-Japan treaty. Since the start of fishery regulations in 1952, the number of vessels (catcher boats) participating in the mothership fishery reached a peak in 1956, and has subsequently declined more or less steadily to 1988 (Figure 3a). Several important restrictions have been imposed on the fleet over this period.

Fig.1. The area of operation for the Japanese landbased fishery

Fig.2. The area of operation for theJapanese mothership fishery

In the early years of INPFC regulation, the fishery was excluded from waters west of approximately 175°E, in order to protect American origin salmon (Gulland 1983). It became clear later, however, that American-origin salmon range considerably further than had at first been thought. The INPFC treaty was subsequently renegotiated in 1978 to limit the numbers of vessels operating. From 1978 to 1986, on average, 72 % of Japanese fishing effort occurred inside the US (EEZ). During this period the problem of the incidental capture of marine mammals became an important factor in the regulation of the fishery.

In the 1960's, several publications (reviewed by Ohsumi, 1975) indicated that large numbers of Dall's porpoise (Phocoenoides dalli) were killed annually in the Japanese driftnet fisheries. After the passage of the US Marine Mammal Protection Act (MMPA) in 1972, it became necessary under domestic US legislation for Japanese fisheries operating in US waters to have permits for the incidental capture of marine mammals. After the renegotiation of the INPFC treaty in 1978, US observers were placed aboard motherships operating in the US EEZ to monitor catch rates. In 1980, US observers were also permitted on some of the catcher boats, and from 1981 observer coverage was more complete, around 2% of operations being observed outside the US EEZ and 6–12% inside the EEZ (Jones 1984, Anon 1988a). From 1981 a marine mammal permit was issued to the Japanese fleet to allow the capture of a certain specified number of marine mammals. From 1981 to 1986 the quota for incidental catches of Dall's porpoises in the US EEZ was set at 5500 annually. In 1987 this quota was substantially reduced (Anon 1988a). Fishing inside the US EEZ was permitted then only until the marine mammal quota had been reached.

In addition to these restrictions, in 1986 the INPFC treaty was modified to phase out driftnetting for salmon in waters north of 56°N by 1994. In 1988 further restrictions were imposed. Firstly the Soviet Union reduced its quota for salmon catches. Soviet quotas had been reduced from 87000 tonnes in 1957 to 3276 tonnes in 1988, a 96% reduction. Secondly, after an injunction was taken out in the US courts by Alaskan fishermen and wildlife conservation groups under the MMPA, a permit to take marine mammals was not issued in 1988 and Japanese driftnet fisheries were therefore effectively banned from the US EEZ. The overall picture of the regulation of the mothership driftnet fishery is therefore a complex one, but the result has been a considerable decline in total fishing effort since the 1950's, as is shown in Figure 3b.

Fishing effort is measured in terms of the number of tans set. A tan is a net panel, the length of which varies depending upon the type of netting and the way it has been rigged. Typically, tans in the salmon driftnet fishery may be in the region of 45–50 metres in length as measured along the floatline. It should be noted that in other fisheries, where different mesh sizes are used and the net is rigged differently, a tan may measure as little as 33 metres. When comparing catch rates within the same fishery, between years or areas, it is therefore appropriate to measure fishing effort in terms of the number of tans set, but when comparing catch rates of non-target species between fisheries which use different types of netting and different sized net panels, it is more appropriate to convert to some other unit of length measurement. In this report, the number of tans used are reported where this information is available, but to facilitate inter-fishery comparisons, some conversions have been made to give approximate lengths of netting deployed in kilometres as well, when this is useful.

Fig.3. Japanese mothership salmon gillnet fishery

Under the terms of the latest INPFC regulations and bilateral USSR-Japan regulations, the salmon driftnet fishery is also subject to a number of time as well as area restrictions. Overall, the fishery operates from 31st May to the 31st July, or until the Soviet quota is filled or the US marine mammal quota is filled, whichever is the sooner.

Since 1988, however, with the effective zero quota set inside US waters, this last stipulation has not applied, although monitoring of the marine mammal take has been continued outside the US EEZ by Japanese observers. The areas fished are shown in Figure 2.

In 1988 43 catcher boats operated in the North Pacific and Bering Sea zones, setting some 482 000 tans between them, which amounts to around 21690 km of netting, over an 8 week period. In Figure 4 the seasonal distribution of effort (number of tans set) by 10 day period is plotted for 1988, and in Figure 5 effort is plotted by 1 degree square for the 1988 fishery. In 1988 the fishery was over by the 20th of July, and was concentrated in the two high seas areas of the North Pacific (170°E to 175°E and 48°N) and the Bering Sea (175–179°E and 56–57°N); fishing in the second of these two areas is due to be phased out by 1994.

The current area of fishing may be compared with that of 1986, prior to the exclusion of the fleet from US waters, in Figure 6. Fishing effort was even more widely dispersed than this prior to the renegotiation of the INPFC treaty in 1978. Peterson (1974) shows catch and effort maps for the mothership fishery in the 1960's, which show that effort was distributed from 47°N to 60°N and from 160°E to 175°W.

The decrease in fishing effort by the mothership fleet which has occurred since the 1950's has therefore been accompanied by a contraction of the range of the fleet, and a contraction in the annual duration of the fishery, from 3 months in the 1970's to 8 weeks at present. Catches of the major target species involved have also declined considerably, and it must be assumed, so too have catches of non-target species.

Current catches of the five target species are given in Table 1, and total catch per unit effort is plotted in Figure 7 for the years 1983 to 1989. Interpreting the CPUE series would be difficult because of the large changes in the fishing area over this period.

The gill nets used by the mothership fishery are made from a nylon monofilament yarn with a stretched mesh size of from 121 to 130 mm. Nets are approximately 8 metres in depth, fishing up to the surface. Two or three net sections may be set, with a maximum total of 330 tans, or roughly 15 km. Catcher vessels operate at least 8 km apart, with nets deployed parallel to one another at a slight angle to the prevailing swell, usually at 20° or 200° in the US EEZ. Nets are set at dusk and lifted around dawn, fishing for around 9 or 10 hours (Anon 1988a).

Fig.4. Japanese mothership salmon gillnet fishery Seasonal distribution of effort 1988 (No of tans set by ten day period)

Fig.5. Distribution of fishing effort in 1988 (thousands of tans set)

Fig.6. Distribution of fishing effort in 1986 (thousand of tans set)

Japanese mothership gillnet salmon fishery, catch in kilograms.

Fig.7. Japanese mothership salmon gillnet fishery Catch per unit effort (tonnes of salmon per thousand tans) 1983–1989

Fig.8. Japanese landbased salmon drift gillnet fishery Fishing effort (thousand of sets made) 1952–1988

The Japanese landbased fishery has been slightly less well documented, in English at least. The number of vessels participating in the landbased fishery has declined since the 1970's. Two types of vessel operate in the landbased fishery: smaller coastal vessels ( < 30GT) and medium sized high seas vessels (< 30GT). During the period 1973–1977 some 1400 coastal vessels were operating, but this was reduced to 678 during the years 1978–88. There were some 330 landbased high seas vessels in 1973–77, but this fell to 209 during the period 1978–86, and has subsequently fallen to 156 in 1989 (Anon 1989, Suzuki 1990). Fishing effort in the landbased fishery is plotted in Figure 8 for the years 1952–87. The general area in which landbased fishery operates was shown in Figure 1. In Figure 9 the fishing effort of the landbased fishery for 1988 is plotted by 1° square. Major fishing areas lie between 43°N and 45°N, and between 157°E and 172°E. The fishery currently operates between late May and the end of June, and fishing effort (standard tans of netting set) by ten day periods 1986–1988 are shown in Table 2, together with catches of the five salmon species. Catches in the landbased fishery have declined in recent years from 20342 tonnes in 1983 to 7,700 tonnes in 1989 (Suzuki 1990). The duration of the fishery has also evidently declined, from May, June and July in 1962 to just May and June by 1988. A spatial shift has also occurred sine the 1960's, and the landbased fishery has most recently been concentrated nearer to the mothership fishery (Anon 1988a).

The nets used by the landbased fishery are similar to those used in the mothership fishery but the mesh size is smaller, from 110–117mm. The vessels are smaller too, and coastal vessels of under 10 GT are limited to 10 km of netting, while the offshore boats may use up to 15 km. Nets are separated by 6 km rather than 8 km for both sectors of the landbased fishery.

The reductions in fishing effort, and declining catches in both the Japanese driftnet fleets which have occurred over the past two or three decades, coupled with current historically high catch levels of salmon in their countries of origin would suggest that the impact which Japanese salmon driftnet fisheries have on their target species may not be high (FAO 1990).

Non-target species

The accidental capture of non-target species in this fishery has been remarked upon, and documented for several decades. Records from the commercial fishery are augmented by records from research vessels employing both the same and different types of gill net.

A wide range of non-target species is included in the catch of the salmon driftnet fishery. These include mammals, birds, fish and squid. No reports of accidental captures of turtles could be found, although the Japan Fishery Agency's numerical codes for species caught in salmon research vessel nets includes a code for Chelonia mydas the green turtle. This is generally regarded as a tropical species, however, and leatherback turtles (Dermochelys coriacea) might be expected more frequently in these waters of the Northern Pacific.

Fig.9. Distribution of effort (thousands of tans set) in the landbased fishery in 1988

Japanese landbased driftnet salmon fishery. Catch by salmon species in numbers.

Of the mammals, the Dall's porpoise (Phocoenoides dalli) is the most frequently caught. Mizue and Yoshida (1965) reported catch rates of 0.88 porpoises per set in 1964. These data were collected on a voluntary basis from the mothership driftnet fleet. Assuming 330 tans of netting per set and 45m per tan, then this figure is equivalent to 59.35 porpoises per 1000 km of netting set.

Ohsumi (1975) reported on the catch rates of porpoises and other cetaceans in research vessel cruises. Among 52 identified individuals, 48 were Dall's porpoises, one was a harbour porpoise (Phocoena phocoena), one was a pilot whale (Globicephala sp.) and two were Baird's beaked whales (Berardius bairdii). Catch rates were, on average, for the period 1962–1971, 0.48 cetaceans or 0.46 Dall's porpoises per 1000 tans, which is equivalent to 10.67 cetaceans or 10.22 Dall's porpoises per 1000 km. Ohsumi noted an increasing catch rate through the year from April to November, and stated that there was no trend in CPUE over the period examined.

Jones (1984, 1990) has summarised more recent information on catch rates (numbers per set) of Dall's porpoise in both landbased and mothership fisheries. These have been converted into catch rates per 1000 km of netting set for comparative purposes, and are presented in Table 3. An interesting feature of these data is the increasing trend between 1978 and 1987 in the catch rates reported from the Japanese mothership fleet. Several reasons might be suggested for this trend. Firstly, the porpoise population may have increased. Secondly, the changes in a real distribution of the fishery described above may have resulted in a greater proportion of total effort being expended in areas of relatively high porpoise density. Thirdly, reporting efficiency may have improved over the period of data collection.

Some support for this latter suggestion comes from the fact that after 1980, when US observer data were used to estimate catch rates independently, US and Japanese estimates of total catch have shown some convergence (Table 4a). Presumably some feedback from the US observer scheme may have improved reporting efficiency in the fleet as a whole.

Ito (1986), also shows an increase in the entanglement rate of Dall's porpoises in Japanese salmon research vessel nets between 1962 and 1984. However, as Ito points out, prior to 1978 entangled animals did not necessarily include those dropping out of the net or released alive, so again, the trend may be more due to changes in reporting procedures. Furthermore, it is not clear whether or not Japanese research cruises have altered the area of their operations during this period, which could also influence porpoise catch rates as suggested above. Ito's catch rate data have been combined with catch rates from Japanese research cruises for 1985–1987, and the combined data sets are shown in Figure 10.

⁺ Data from Jones (1984) converted from numbers per 1000 sets,assuming 1 set consists of 330 tans, where a tan is here reported to have a ‘standard’ 50mlength
^*US EEZ data except where otherwise indicated.
¹18 driftnet operations observed by US observers
²10 operations - US research vessel
³Range of catch rates from US observers in US EEZ

+ On the basis of US Observations

4b. Estimated total catches in numbers for the land based fishery.

Estimate 1: Based on an assumed catch rate of 0.22 per 1000 tans (from Ito 1986)

Estimate 2: Based on the observed catch rate in the Mothership fishery.

Estimates of the total catch of Dall's porpoises have been made both by extrapolating from estimates of the catch rate, and from reported catches in the mothership fishery. Jones (1990) provides a table of estimated catches based on catch rates and on the reported catches in the mothership fishery, 1981–1987. These data are shown in Table 4a, where a convergence in the two estimates can be seen. Ito (1986) points out that the mean rate of entanglement in the US observer data set is 0.47 porpoises per set (equivalent to 31.65 per 1000 km) whilst the mean catch rate in the Japanese salmon research vessel cruises was almost identical at 0.46 porpoises per set (equivalent to 30.98 per 1000 km). This agreement suggests that estimates of total catches based on observed catch rates may be more reliable than those based on catch records.

Catches in the landbased fishery are less easily estimated. Reported catches are given in Table 4b together with estimates of catches based on the catch rate observed in the mothership fishery (from Jones 1990). Ito (1986) considered that on the basis of Japanese research cruises, catch rates in the landbased fishery were lower than those in the mothership fishery. Ito reported an average catch rate of 0.2 porpoises per set in the area of the landbased fishery (13 per 1000 km). Estimates of porpoise catches by the landbased fishery based on these lower catch rates are also given in Table 4b.

It is clear from the numerous studies which have been conducted on the interaction between Dall's porpoise and the salmon driftnet fishery that seasonal and distributional factors may influence catch rates, although a clear understanding of the factors involved has not yet been obtained (Ito 1986).

Catches of other marine mammals have been less well studied. Smaller catches of northern fur seals (Callorhinus ursinus), harbour porpoise (Phocoena phocoena) and Steller's sea lion (Eumatopias jubatus) have been reported (Anon 1988a). O'Hara et al (1986) state that 400–1000 fur seals are taken every year, and add northern right whale dolphins (Lissodelphis borealis), Pacific white-sided dolphins (Lagenorhynchus obliquidens) and common dolphins (Delphinus delphis) to the list of species sometimes caught. Ohsumi (1975) reported pilot whales and Baird's beaked whales. Quantitative data on such catches are rarely published. Catches in Japanese salmon research cruises are given in Table 5. Additional species include Curvier's (goosebeaked) whale (Ziphius cavirostris) and killer whale (Orcinus orca). Northern fur seal catch rates were just less than a half those of Dall's porpoises, but it should be remembered that the use of the gear, and area of operation in research vessels may be very different from the commercial fleet. Numbers of other mammals caught are so low as to make any useful estimates of catch rates impracticable. Reported catches in the mothership fishery during 1985–1986 were limited to 11 northern fur seals (6 of which were alive) and one harbour porpoise.

Fig.10. Catch rates of Dall's porpoise in Japanese salmon gillnet research cruises

Catches of marine mammals in Japanese salmon research cruises

A : Ohsumi (1975).
B : Ito (1986).
C : FAJ (1986).
D : FAJ (1987).
E : FAJ (1988).
+ Pinnipeds not recorded.
* Columns B - E only.

At least twenty three species of seabirds are known to be caught in the salmon driftnet fisheries of the North Pacific (Jones and DeGange, 1988). Seabird catches have been studied by King et al. (1979), on the basis of research gillnet operations in the mothership fishery area in 1978, by Ogi (1984) based on Japanese salmon research cruises between 1977 and 1981, by Ainley et al (1981), based on research gillnet operations in the mothership fishery area in 1979, by Sano (1978) based on salmon research cruises in the landbased fishery area in 1977, and by DeGange et al. (1985), based on US observer records from the Japanese mothership fleet inside the US FICZ. These studies are reviewed by Jones and DeGange (1988), and are summarised briefly below.

Bird mortalities in the landbased fishery during 1977 were estimated by Sano (1978) to be between 113,000 and 232,000 (all species). Effort in the landbased fishery has subsequently declined, so that catches may also have decreased. Jones and DeGange (1988) provide a more recent estimate of the numbers of birds killed in the landbased fishery. Their figures are reproduced in Table 6. These estimates were based on catch rates observed in 1977, but the year to which the estimated catches apply was not stated.

Bird mortalities in the mothership fishery were estimated by King et al. (op cit), to number 75,000 to 250,000 in total. Ainley et al (1981) estimated 266,500 in total, and Ogi (1984) estimated a range of from 130,000 to 180,000 between 1977 and 1981. These estimates, however, were based on research cruises which may use gill nets of different selectivities from the commercial gill net operators. Ainley et al adjusted for this potential bias. More reliable results should be obtained from observations of entanglement rates on commercial vessels. Jones and DeGange (1988) have used observed catch rates from the commercial fishery inside the US EEZ to estimate total catches for the years 1981–1984. Their results are reproduced in Table 7.

Major catches are of sooty and short-tailed shearwaters (Puffinus griseus & P. tenuirostris), and also of several alcid species; both of these groups are ‘pursuit’ feeders, and they are generally caught in the top two metres of the net. Scavengers and surface feeders are rarely caught (King 1984).

Very little information has been published on the capture of non-target fish and squid in the salmon driftnet fishery. It would appear that the only available information consists of the data records for fishes and squids caught incidentally in gill nets of Japanese salmon research vessels, which are submitted annually to the INPFC by the Fishery Agency of Japan. Records for 1988 are summarised in Table 8.

Apart from 8 species of salmonids, some 52 species of fish and squid were recorded during the 1988 research cruises. Most of these were caught only in very small numbers. Certain species, however, were caught relatively frequently. Pacific saury were caught in large numbers, as were Ray's bream, scorpionfish, walleye pollack and three squid species. The large numbers of sardines caught; however, should warn against using these research cruise data to extrapolate to the entire fishery. As this species of sardine reaches a maximum length of 25cm (Kamohara 1967) it seems unlikely it would be caught in such large numbers in salmon gill nets of 110– 130mm mesh size. Rather, these catches reflect the range of mesh sizes used by the research vessels. As indicated previously, therefore, catch rates in the research gill nets do not necessarily reflect those of the commercial fishery.

^a From DeGange, Forsell and Jones (1985). Data collected in 1977.

^a Mean calculated by averaging across four years.

Summary of fish and squid species caught in Japanese salmon research nets in 1988.

2.1.2 Squid driftnet fisheries

The driftnet fisheries for squid in the North Pacific involve three nations, namely Japan, Taiwan (Province of China) and Korea. Driftnetting for squid was initiated by Japanese vessels in the autumn of 1978 in northwest Pacific waters. At that time fishing was restricted to the waters off eastern Hokkaido and Honshu, in waters west of 150° E. Fishing efficiency was found to be greater than jigging, and this resulted in some conflict between the driftnetting and jigging fleets. As a consequence the Japanese government restricted the driftnet fishery to waters north of 20°N and west of 170°E after January 1st 1979 (FAJ 1982a).

Despite this regulation, the fishery continued to expand, and in 1981 additional measures were introduced to control Japanese fishing effort in this fishery. Under this regime the Japanese government limits the number of vessels approved to fish for squid with driftnets and limits fishing to within the area enclosed by 20°N, 46°N, 145°W and 175°E. In addition to these aerial restrictions, the northern limit of permitted fishing is moved throughout the fishing season to limit catches of salmon in the more northerly areas of the fishing zone. The regulations for fishing at the northern limit of the fishing zone are summarised in Table 9.

Two classes of vessel are recognised in the Japanese squid driftnet fleet, these are from 60 to 100 GT and from 100 to 500 GT. There are also two types of government approval for vessels driftnetting for squid, the one permitting 7 months of operation from June 1 to December 31st, and the other allowing 4 months of operation from August 1st to November 30th. The numbers of vessels operating in the Japanese fleet over the period 1981 to 1989 are shown in Table 10. Vessels operate fishing trips of from 30 to 70 days in general, and freeze their catch on board. Landing is permitted at 37 designated ports in Japan, but each vessel can land at no more than three chosen ports. Landing at other ports or trans-shipment at sea is not allowed. The capture of salmon and trout is also not allowed, and vessels and their nets must be marked in accordance with the Fishery Agency's regulations. Fishing effort in 1988 is shown in Figure 11.

The gear used is constructed from a nylon monofilament yarn No. 9 or No. 10 (about 0.5mm in diameter). Mesh sizes are regulated and must be greater than 100mm and not more than 135mm. Nets are 9 or 10 metres in depth, and are fished up to the surface. Typically mesh sizes of 115mm to 120mm are used. 450 to 1,100 tans of netting may be set per day, but generally, for reasons of operational efficiency between 450 and 900 tans are set per day. Each tan has a fully extended length of between 72m and 90.9m, but the operational length after rigging is only about 46% of this (FAJ 1985), so that, in this fishery a tan may be between 33 and 42 metres in length when it is being used. The ration given (0.46) can also be termed the horizontal hanging ratio (E₁). Fridman (1986) states that driftnets are generally constructed with a hanging ratio in the 0.5 to 0.7 range, whilst values as low as 0.3 might be used to increase tangling. The hanging ratio employed in the Japanese driftnet fishery therefore, may have a marginal tangling effect as well as its gilling function.

a - Only large mesh (tuna) nets
b - Was moved by Japanese from 42N to 43N in 1988
c - Were both moved north by Japanese from 44N in 1988
d - Taiwanese regulations restricted these fisheries to 39N in 1985

N.A. indicates data are not available

Fig.11. Distribution of Japanese squid driftnet fishing effort (no. of tans set) in 1988

The amount of netting deployed can be compared with that in other fisheries by converting the approximate number of tans used into metric units. Assuming that between 450 and 900 tans are set by each vessel per day, as stated above, and that in this fishery each of the tans is between 33 and 42 metres in length, then each vessel would set between 14.85 and 37.8 km of netting per day. Tsunoda (1989), however, reported that the Japanese vessel on which he worked as an observer in 1986 deployed 24 mm of netting, equivalent to 44.45 km. Observer reports from the 1989 joint Japan-US-Canada observer program indicate that, from 1402 observed gill net sets, the average number of tans deployed was 1186, which should be equivalent to between 39 and 50 km of netting, depending on the length of a tan in this fishery. Nets on Tsunoda's vessel were set in the late afternoon, an operation which took up to 3 hours, and hauled 3.5 to 4 hours before sunrise, an operation which took 6.5 to 10.5 hours after a minimum 6.5 hour fishing period.

Korean vessels joined the squid driftnet fishery in 1979 (3 vessels) and increased in number rapidly to 1984 (Table 10). Most of the vessels in the Korean fleet are around 350 gross tons, but at least 11 may be more than 400 tons (Gooder 1989). Korean authorities did not restrict the time or area of fishing by Korean vessels until 1989, but incidentally caught salmonids must be discarded (Anon 1989). Korean restrictions imposed since 1989 are shown in Table 9. The fishing season starts in late April in an area 36°N to 43°N and 160°W to 170°W, and continues there until late July or early August; from early August to mid-December smaller squid are caught in the “eastern grounds” located between about 36°N and 43°N and between longitudes 143°E and 169°E (Gooder 1989).

Korean vessels employ nets with mesh sizes of 86 to 155mm, although most are said to be in the range of 96 to 155mm (Anon 1988a). Gooder (1989) found that, on the vessel he observed, the larger squid on the western ground were fished with nets primarily of 105mm mesh size, with some 115mm meshed nets, whereas on the eastern ground most of nets were of 86mm mesh, with some 96mm meshed nets. Nets are approximately 9m in depth, and are set in units of ‘poks’, panels of netting, equivalent to Japanese tans, which are reportedly around 50m in length. The number of poks deployed by each vessel per set is reported to have increased from around 200 in 1980 to 540 in 1983 and 1000 in 1987 (Anon 1989). This was considered an operational maximum amounting to about 50 km of netting per vessel, assuming 50m per pok.

Gooder (1989), however, reported that the vessel on which he worked as an observer, which was 430 gross tons (larger than average) set 1400 poks per operation, which would be around 70 km. Gooder noted, however, that there was a major discrepancy between the estimated length of the net, and the distance steamed between the beginning and end of a set, as measured by dead reckoning, which was apparently of the order of 40 to 50 km.

The Taiwanese fleet operating in the North Pacific driftnet fishery increased from 12 vessels in 1980 to 166 in 1988 (Table 10). Most of the Taiwanese fleet averages around 390 GT (47m). Information on the Taiwanese fleet is sparse. Monofilament nylon nets are said to be used, with meshes of 94mm, and the number of tans per vessel is said to have increased from 250–500 to over 1000 during the 1980's. The fishing season is generally from May through October, and fishing grounds have been extending westwards through the Pacific in recent years (Anon 1988a).

Fig.12. Catches of squid (in tonnes) by the three squid driftnetting nations 1978-88 (Catch data are incomplete for Taiwan (Province of China) and Korea)

Statistics on catches of squid by the three fleets are not widely published. Some of the available data are summarised in Table 10, and are shown in Figure 12. Japanese catches dominate the total.

Some interesting features emerge from Table 10. Firstly, catches per boat vary widely between the three fleets. This is almost certainly partly due to differences in the length of the fishing seasons, Korean vessels spending up to 175 days fishing on average in 1986, whereas Japanese vessels fished for only 72 days on average in the same year. Data on fishing effort by the Taiwanese fleet could not be found. There is no obvious trend in catch per day in the Japanese fleet between 1983 and 1988, but catch per tan by the Korean fleet may have declined, although the data are sparse.

Catches of Non-Target Species

Following concerns during the 1980's that the squid driftnet fisheries might be impacting marine resources, notably salmon, originating in US waters, the US government passed the Driftnet Impact Monitoring, Assessment and Control Act of 1987. This act stipulated that the US should negotiate with each foreign government which “conducts, or authorizes its nationals to conduct, driftnet fishing that results in the taking of marine resources of the United States in water of the North Pacific Ocean outside of the exclusive economic zone and territorial sea of any nation, for the purpose of entering into agreements for statistically reliable co-operative monitoring and assessment of the number of marine resources of the United States killed and retrieved, discarded or lost by the foreign government's driftnet fishing vessels”. Failure to enter such an agreement would lead to the certification of the offending nation under the Fisherman's Protective Act 1967.

Following the passage of the Driftnet Impact Monitoring, Assessment and Control act, observer programs were established between all of the nations involved. Prior to 1989 individual squid driftnet vessels had been observed by US scientists (Cary and Burgner 1983, Ignell et al 1986, Tsunoda 1989, Gooder 1989). Japanese scientists also conducted an observation program on 10 driftnet vessels in 1988 (FAJ 1989a). In 1989 the first joint observer program was initiated and detailed catch data were obtained (Gjernes et al. 1990).

In addition to these observations made on board commercial vessels, data have also been collected by US, Canadian and Japanese scientists on catches of non-target species in research cruises. Japanese fishermen are also required to report catches of marine mammals, and these data on reported catches are submitted annually to the INPFC.

Japanese squid research cruises employ a range of mesh sizes, much smaller and much greater than those used in the driftnet fishery. As with the salmon research cruises referred to above, it is therefore somewhat difficult to interpret data on catches of non-target species in relation to the commercial fishery.

There is some indication that the ‘non-commercial’ nets used in research cruises, and most notably larger meshed nets tend to catch marine mammals more frequently than the standard ‘commercial’ (115mm) mesh sizes. For example, Murata et al (1988), reported 8 northern fur seals, and 2 Dall's porpoises taken in 45 driftnet operations in 1987. A total of 15 different mesh sizes were used in each operation ranging from 33 to 197mm. All the marine mammals were taken in the 197mm mesh, except for 1 fur seal taken in a 157mm mesh net section.

Further investigation on marine mammals catches in research cruises is reviewed below briefly. Murata (1987) reported a total of 16 fur seals and 2 Dall's porpoises taken in 42 gill net operations in 1986. The porpoises were taken in 110mm mesh net sections, fur seals were taken in meshes of 48mm, 63mm, 110mm (x3), 115mm (x4), 121mm, 157mm (x3), and 197mm (x7).

Kawasaki et al (1989) reporting on a 1989 cruise in which 29 sets were made, half with commercial (115mm) mesh and half with a range of experimental size meshes. Four fur seals were taken, 3 in ‘commercial type’ nets, and 1 in a ‘research’ net, 3 Pacific white-sided dolphins were taken, all in ‘research’ nets, as well as 18 shearwaters, 16 of which were in ‘commercial type' nets.

Yatsu (1989) provides data on animals caught in 33 gill net operations in 1989; a total of 1650 tans (ca. 82.5km assuming 50m per tan) of research net and 1650 tans of ‘commercial type’ nets were set. Six marine mammals were taken, all of them in the research nets. Yatsu's results are reproduced in Table 11. Three species of bird and 13 species of fish and squid were also identified.

Murata et al (1989) record the capture of 5 northern fur seals, 1 Dall's porpoise, 2 common dolphins and 2 striped dolphins in 3 research cruises in 1988 with a total of 73 operations involving 7260 tans (c. 363 km) of netting. Of these 10 marine mammals, one fur seal and one common dolphin were taken in meshes of less than 100mm, 2 striped dolphins were taken in 115mm gill nets, and the remaining six were taken in larger meshed gill net sections (121mm x 2, 138mm and 157mm x 3).

It is not possible, on the basis of these data alone to estimate the relative efficiency of the various mesh sizes in entangling marine mammals, but larger meshes would appear to be most frequently involved. These reports also give some indication of the frequency with which marine mammals may be taken in driftnets in the area of the squid fishery. A summary of the research cruise results for 1986-88 is presented in Table 12.

If it is assumed that each tan is equivalent to 50m of netting then an approximate entanglement rate can be found in terms of mammals per 1000 km of netting set. The actual length of an experimental tan is not reported, and some commercial tans may be as short as 33 metres, so the entanglement rates are conservative values. It should be remembered that these results are not strictly comparable with the commercial fishery because somewhat more than half the netting employed different mesh sizes. Although the amounts of netting deployed were not large, catch rates appear to be higher than most of those observed in the commercial fishery (see below).

Canadian research cruises have been conducted over several years using driftnets to examine the potential interception of salmonids in the squid fishery. Results are reported by Bernard (1986), LeBrasseur, Riddell and Gjernes (1987), LeBrasseur et al (1988) and McKinnell et al (1989). Standard 50m tans of 115mm mesh monofilament have been used throughout. Marine mammal catches, as reported in these four cruises are summarised in Table 13.

(Figures in parentheses are approximate entanglement rates per 1000 km of netting for comparative purposes)
¹Yatsu (1989).
²Murata et al (1989).
³Kawasaki et al (1990).
⁴Murata et al (1988).
⁵Murata (1987).

*released alive.

Jones (1988) has reviewed US observations on entanglement in 1986 and 1987. Her figures are reproduced below in Table 14.

(Figures in parenthesis represent the numbers of fur seals actually killed)

Experimental fishing trials for squid have also been conducted, by the Canadian Department of Fisheries and Oceans. Six experimental operations were conducted between 1979 and 1988. Flying squid were found to be present in Canadian waters in concentrations sufficient for commercial fishing. In the first four trials few marine mammals were taken (3 in 1983 and 6 in 1985). In 1986, however, a total of 50 mammals were entangled, 3 were released alive. In 1987, 90 mammals were taken. Catches of mammals are shown in Table 15, from Jamieson and Heritage 1988. Catch rates are shown, again in terms of individuals per 1000 km of netting, for comparative purposes. This fishery was terminated in 1987 due to the high mammal catch rates.

Data from two additional observers are available. Tsunoda (1989) made observations on a Japanese driftnet vessel in 1986. Mammal catches consisted of 43 northern right whale dolphins (2 alive), 8 Pacific white-sided dolphins, 7 Dall's porpoises (1 alive), 16 northern fur seals (2 alive) and 1 striped dolphin, 30 sets were observed, each of approximately 44 km.

Gooder (1989) made observations on board a Korean squid driftnet vessel in 1988. A total of 22 sets were reported, each employing about 70 km of netting. Entanglements included 15 northern right whale dolphins, 1 Dall's porpoise and 2 northern fur seals. Both fur seals were released alive.

In 1988, 11 Japanese observes were placed on board 10 Japanese squid driftnetting vessels and observations were made on 464 operations involving 533,618 tans standardised to 50m in length each, equivalent to 26,681 km of netting. Non-target species recorded included northern fur seals, Pacific white-sided dolphins, northern right whale dolphins and Dall's porpoises. Results are shown in Table 16.

Finally, observations have been made in 1989 on 32 Japanese vessels under the co-operative observer scheme established by bilateral agreements between Japan and Canada and Japan and the USA. This Observer program is continuing in 1990 and 1991, with observers on board Taiwanese and Korean vessels too. Results of the 1989 observations have recently been made available (Gjernes et al 1990). A summary of reported catches is shown in Table 16. A total of 1,427,225 tans were observed (1402 operations). This set of observations therefore, represents by far the largest observational data set for this fishery so far.

The various assessments of marine mammal entanglement rate have been summarised in Table 17 for the 5 most important marine mammal species. Clearly many of these observations involved relatively small amounts of netting so that equal weight should certainly not be given to all of them. The most recent data from the 1989 joint Japanese-US and Japanese-Canadian observations are the most thorough. Nevertheless, there are likely to be changes in entanglement rate between years and between areas, so that until a more comprehensive analysis of several years of joint observer data becomes available and inter-seasonal and regional variability can be accounted for, comparisons between previous observations may be useful. Previous observations do at least confirm that right whale dolphins are relatively frequently entangled, as are Pacific white-sided dolphins.

The waters of the North Pacific in the squid fishing area were unusually warm in 1989 so that some differences between catch rates in 1989 and those in other years might be expected. It is interesting to notice that not all cruises have reported entanglements of right whale dolphins more frequently than other species, and that some have reported much higher rates of fur seal and Dall's porpoise entanglements. Japanese observations in 1988, involving over 26000 km of netting, demonstrate comparatively higher entanglement rates for the more northerly distributed fur seals and Dall's porpoises and comparatively lower rates of entanglement for northern right whale dolphins and Pacific white-sided dolphins.

Entanglement rates can be used to provide some crude estimates of total entanglements in the fishery. To do so requires some idea of the total amount of netting deployed by the fishery. As mentioned above, there are clear differences between the behaviour of vessels from the different nations involved. For example the average catch per boat between 1980 and 1986 for Taiwanese, Japanese and Korean vessels was 173, 310 and 518 tonnes respectively. In part, this reflects the differences in the lengths of respective fishing seasons, but it may also reflect differences in average vessel size and amounts of net deployed.

In 1988 the Japanese fleet is reported to have set a total of 36,055,567 tans of netting equivalent, at 50m per tan, to around 1.8 million km, or around 3,900 km of netting per vessel, for the entire season, on the basis of 463 vessels. The amounts of netting deployed by the 141 Korean vessels and 166 Taiwanese vessels are unknown. If it is assumed that the differences in the reported catches per boat are indicative of differences in the amount of fishing effort, then Taiwanese vessels might set 0.56 × 3900 km of netting per vessel per season and Korean vessels 1.67 × 3900 km per season. A very crude estimate of the total amount of netting deployed on this basis would be around 3 million km of netting in the 1988 season. Very approximate estimates for the total catch of marine mammals in the 1989 season, on the assumption of the same level of effort, and using the 1989 joint observer scheme figures for entanglement rate, have been made and are also given in Table 17, rounded off.

On the basis of the assumptions listed above, at least 2200 northern fur seals, 19000 Northern right whale dolphins, 6000 Dall's porpoises and 11000 Pacific white-sided dolphins might be expected to have been entangled in the squid driftnet fishery in 1989. It should be stressed that these figures are entirely dependent upon use of the observed entanglement rates in 1989 which are not necessarily representative of the fleet as a whole.

A variety of birds are also reportedly entangled in squid driftnets. Jamieson and Heritage (1987, 1988) reported albatrosses, sooty shearwater, slender billed shearwaters, storm petrels, terns and auklets and other alcids. Gooder (1989) reported dark shearwaters, black-footed and Laysan albatrosses, and a horned puffin. Tsunoda (1989) also reported black-footed and Laysan albatrosses, and sooty, flesh-footed and short-tailed shearwaters. Murata et al (1989) included horned puffins, tufted puffins and thickbilled murres. In Table 18, reported catch rates for 20 species are listed, by year and by author. Again some variability is observed between years and observers, but again, the 1989 joint observer scheme provided the most comprehensive coverage in any one year. Results from this scheme are therefore used to provide very approximate totals of numbers of seabirds likely to have perished in squid driftnets in 1989, assuming that 3 × 10⁶ km of netting were deployed and reported catch rates apply evenly to all of this fishing effort.

¹Results of joint observer report (Gjernes et al 1990).
²F.A.J. (1989a) observer report.
³Gooder (1989).
⁴Jamieson and Heritage (1988).
⁵US survey results - from Jones (1988).
⁶US survey results - from Jones (1988), (includes Tsunoda, 1989)

Note: Canadian and Japanese research cruise data are not included due to either very low effort (< 50 km of netting) or different mesh sizes.

A Murata et al. (1989)
B Jamieson and Heritage (1987, 1988)
C Japanese observers
D Gjernes et al. (1990)
E In numbers; derived from col D, assuming 3 × 10⁶ km of nets set in all.

Turtles are reported infrequently in driftnets, and unfortunately are rarely identified to species level. Three unidentified turtles were observed entangled by FAJ observers in 1988 in 464 observed operations involving 27000 km of netting (0.1 per thousand km). Gjernes et al (1990) identified 9 leatherbacks, 1 green turtle and 1 olive ridley among 22 entangled turtles, the remaining 11 being unidentified. This represents a higher entanglement rate (0.31 per 1000 km). If the higher rate is taken as representative of the fishery in 1989 then approximately 900 turtles may have become entangled in that year. At least 6 of the 22 were leatherbacks which drowned (3 other leatherbacks were released alive). If only 27% of all turtles caught were leatherbacks which drowned, then some 250 leatherbacks might be expected to have drowned in 1989. If more of the unidentified 13 turtles were also leatherbacks, then this leatherback total could be higher.

As with the salmon driftnet fishery, a wide variety of fish are caught in squid driftnets. The most numerous of non-target species is the Pacific pomfret. Blue Sharks are also taken in large numbers, as well as a variety of tunas. Catch rates for some of the fish species are given in Table 19, together with estimated total catches based catch rates reported by Gjernes et al (1990).

These estimates for total catches of fish, turtle, bird and mammal species should be considered as initial approximations only. Clearly a more detailed analysis of catch rates by area and by month or year, coupled perhaps with oceanographic data, might help to explain some of the variability observed in preceding tabulations of observed catch rates. Confidence limits might then be applied to some of the estimates. In the meanwhile the figures proposed above may serve to illustrate the possible scale of some of the catches, and as such are discussed again in Section 3 below, in relation to known population sizes and impacts on the species involved.

2.1.3 Large mesh driftnet fishery

The third major driftnet fishery in the North Pacific is the large mesh driftnet fishery, for tunas and billfish which has, at least until recently, been largely Japanese. The origins of this fishery have been traced to a coastal driftnet fishery which began off the Pacific coast of Japan in the 1840's, targeting bluefin tuna (FAJ 1982b). Motorised vessels were introduced in 1914, and the fleet is reported to have expanded to over 250 vessels by the 1930's, and subsequently to have disappeared in 1940 after the decline of the bluefin tuna stock (Anon 1988a). After the Second World War the fishery was apparently revived and continued to operate largely in coastal Japanese waters until the 1970's. Separate statistics on catches by this fleet are not available, but it is believed that from 1952 to 1972 the large mesh driftnet landings amounted to only a few hundred tonnes per year (Anon 1988a).

During the 1970's the fishery expanded from Japanese Pacific coastal water to other coastal waters of Japan, and to offshore areas including the South China Sea and Yellow Sea. The range of target species was also expanded to include marlin, skipjack and other tunas. In the late 1970's and early 1980's the fishery expanded further east in the Pacific and onto albacore stocks. The fishing area in the North Pacific is shown in Figure 13. From the mid 1980's the fishery also expanded into the South Pacific (see below).

The fishery is seasonal, operating largely between January and May (Suzuki 1990). However, marlins and tunas are taken off the Sanriku district of Japan from June to October, and also in the South China Sea and the Yellow Sea from August to December (FAJ 1982b). Most of the Japanese vessels in this fishery are also used in other fisheries, and in the early 1980's about half were said to make only around one trip per year, of up to 100 days, in large mesh operations (FAJ 1982b). At that time the fleet was said to be composed largely of vessels in the 60–100 GT range. Suzuki (1990) states that, more recently, vessels have been in the 100 to 500 GT range. During the 1980's Taiwanese vessels have also participated in this fishery, but the numbers of vessels involved is largely unknown. There were reportedly 123 Taiwanese vessels in 1989, operating mainly in the months of June to September. There is also a small US fleet operating large meshed driftnets in the eastern North Pacific.

During the 1970's, when the first major expansion of driftnetting began, disputes arose between Japanese squid jiggers and driftnetting vessels in general. There was also perceived to be some potential conflict between the driftnet vessels and other Japanese and foreign tuna fisheries. Regulation of the fishery was, therefore, initiated in 1973.

The Japanese large scale driftnet fishery is now regulated by seasonal and area closures, by gear regulation and by other operational controls. In order to minimise potential and actual conflicts with other fisheries, the fishery is prohibited in Japanese Pacific coastal waters, and in other sensitive areas, and is subject to seasonal closures in other areas (see Figure 14).

Fig.13. Japanese large-meshed driftnet fishery - area of operation

Fig.14. Areas closed to Japanese large-meshed driftnet fisheries up to 1990

Gear regulations include the prohibition of mesh sizes of less than 150mm, and a maximum amount (length) of netting which is permitted to be stowed on board (30 km) or set in the water (12 km). Double sheets of netting (tangle nets) are also prohibited. Gear in use, and driftnetting vessels, must be marked according to the regulations of the Fisheries Agency (FAJ 1982b).

The type of netting employed is most usually nylon monofilament, with a mesh size of 170– 180mm and a 1.2mm diameter yarn (multifilament nets may also be used (Shima 1984)). Fully extended each tan measures around 75m, but they are rigged with a primary hanging ratio (E₁) of from 0.44 to 0.48, so that the length when deployed is 33–36m per tan. Again, as with the squid driftnet fishery, such nets may also have some tangling tendency, as drift gill nets are more usually rigged with a primary hanging ratio of 0.5 to 0.7 (Fridman 1986). Total length of the deployed net is limited to 12 km, but vessels operating in the South Pacific were reported to use 40 km of netting (Coffey and Grace 1990).

The number of vessels involved in this fishery is not well known. Suzuki (1990) provides estimates of the numbers of Japanese vessels involved in this fishery from 1983 to 1988. A peak of 717 vessels was reported in 1982 (Anon 1988a). The reasons for the decline in vessel numbers are not clear, but may be related to the increasing proportion of larger vessels referred to above. Estimates of the numbers of vessels involved in this fishery are reproduced in Table 20 from Suzuki (1990) and FAJ (1982b).

These numbers are estimates and include ‘small scale’ vessels, presumably operating in permitted coastal areas. Catches per boat show an increasing trend overall, through 1987 appears to have been a poor year.

The Taiwanese large-mesh driftnet fishery remains largely unknown. Reports from Japanese pole and line fishermen in the Northern Pacific albacore grounds, published in Japanese trade journals, have indicated that Taiwanese vessels are operating in the area (Anon 1988a). Enquiries to Taiwan (Province of China) on the nature and scale of this, and other Taiwanese driftnet fisheries, elicited no response. Chen (1985) reports that Taiwanese large mesh driftnet fisheries for marlin and sailfish use blue multifilament yarn with a primary hanging ratio of 0.55–0.60 on the upper side and 0.75–0.80 on the lower side.

Catches by the Japanese large mesh fishery are shown in Figure 15 (from Watanabe 1990). There has been an increasing proportion of ‘tunas’ and ‘skipjack’ in the catch, while albacore catches have formed a fairly stable minor proportion of the catch since 1981.

The US large mesh driftnet fishery has operated both within and outside the US EEZ; some 309 vessels from California, Washington and Oregon were reportedly involved in this fishery in 1988 (Douglas 1989). Although the number of vessels which were fishing beyond the 200 mile limit is not known, “some” have been reported, in the past, to switch from shark and swordfish fishing to target albacore outside the US EEZ seasonally (Anon 1988a). More recently Barlow et al (1990) report that the fishery is currently in decline, supporting only around 150 vessels fishing within the US EEZ, and that fishing off the Oregon and Washington coasts has been discontinued due to high levels of non-target catch.

The Californian fishery is described by Miller et al (1983) and Barlow et al (1990). It was initiated in 1977, expanded rapidly, and was initially targeted on sharks. Swordfish were also taken and became an important bycatch. Albacore fishing outside the EEZ is reported to have employed nets with a mesh size of 178–205mm (Anon 1988a). Since November 1990, US fishermen have been prohibited from using large scale driftnets both within the US EEZ and on the high seas.

Catch data from Taiwanese vessels and from US vessels fishing outside the EEZ could not be located. Data on US swordfish and shark driftnet fisheries with the EEZ are discussed further below.

Fig.15. Landings by species group from the Japanese largemesh driftnet fishery

Non-target species

Almost nothing is known about non-target captures in the large mesh fishery. There is clearly a wide range of fish species which are targeted, including sharks, marlins, sailfish and a variety of tunas. O'Hara et al (1986) report that a Japanese gill netter was found illegally fishing inside the US EEZ around Hawaii, and when boarded the vessel's log indicated that it had taken 69 cetaceans. Watanabe (1990a) reported on cetaceans taken in large mesh driftnet fishery surveys in the North Pacific in 1982/3 and in 1989. In 51 such operations in 1982 and 1983, 22 right whale dolphins, 28 striped dolphins, 16 common dolphins, 4 Risso's dolphins, 2 spotted dolphins, a bottlenose dolphin, a Pacific white-sided dolphin, a pygmy killer whale and an unidentified beaked whale (total 76 cetaceans) were taken. In 1989 15 operations were conducted and 2 pygmy sperm whales and a Pacific white sided dolphin were taken.

The US Driftnet Impact Monitoring, Assessment and Control Act of 1987 requires the Secretary of Commerce to institute monitoring of this fishery and observers have been placed on board Japanese and Taiwanese vessels in 1990. This programme is scheduled to continue in 1991.

Japanese research on selectivity of large meshed gill nets suggests that the commonly used 160mm and 170mm meshes are unlikely to gill salmon (FAJ 1982b). Studies in the squid driftnet area by Japanese research vessels with large meshed nets, however, suggest that large meshed nets may be more likely to ensnare marine mammals (see above). Certainly the twines used in this fishery are considerably stronger than those used in the salmon or squid fishery. At present, however, the non-target (i.e. non-fish) captures are a matter for conjecture.