4. BIOLOGY AND FISHERIES HISTORY OF THE COMMERCIALLY HARVESTED SPECIES

4.1 Antarctic krill (Euphausia superba)
4.2 North Pacific krill (Euphausia pacifica)
4.3 Other species of krill harvested in Japanese waters
4.4 Krill in the North Atlantic (Meganyctiphanes norvegica, Thysanoessa raschii and Thysanoessa inermis)
4.5 Other potential krill fisheries

Only two species of krill are being commercially harvested to any significant degree and we will concentrate on the biology of these species - E. superba and E. pacifica - as it relates to these fisheries. In the case of E. pacifica, we will concentrate on the Japanese fishery which is by far the largest fishery on this species and is the one with the greatest amount of available information, however much of this information has not yet been made available outside the Japanese literature. There is a considerable amount of existing information on the biology of and the fishery for Antarctic krill so we will concentrate on recent developments for this species and the reader is referred to earlier reports for a more detailed approach. The biology of the other species that are currently being harvested, and of those that have been suggested as having commercial potential will be dealt with in less detail. Basic information on all these harvested species is summarised in Table 1.

4.1 Antarctic krill (Euphausia superba)

4.1.1 Distribution
4.1.2 Life history
4.1.3 The fishery
4.1.4 Management
4.1.5 Ecosystem interactions
4.1.6 Ecosystem monitoring
4.1.7 Bycatch

4.1.1 Distribution

E. superba has a circumpolar distribution, generally being found south of 50°S and away from truly coastal waters in the Antarctic where another, smaller, species, Euphausia crystallorophias, dominates (Fig. 2). Their distribution is largely known from studies in summer when the sea ice has retreated. Their exact winter distribution is uncertain but most of the population can be found under the seasonal pack ice. The exception to this generalisation is the population of krill found around South Georgia, normally north of the pack ice zone, which has sustained a winter fishery in recent years (Everson and Goss 1991). The characteristic feature of the distribution of krill on all scales is its patchiness (Marr 1962). There is also considerable inter-annual variation in the abundance of krill in particular areas and even some evidence of krill being absent from large areas in certain years with dire consequences for land-based krill feeders (Heywood et al. 1985). There is also some suggestion of long-term declines in krill abundance and recruitment, particularly in the South Atlantic (Siegel and Loeb 1995).

Antarctic krill are usually found in the top 200m of the water column but there are many records of the presence of surface swarms of krill (Marr 1962) and of aggregations of krill near the bottom at up to 480m depth (Gutt and Siegel 1994). Antarctic krill, like many species of euphausiid, may undergo diurnal vertical migrations, though E. superba does not exhibit this behaviour in a routinely predictable fashion (Miller and Hampton 1989a).

Although Antarctic krill are only found throughout the Southern Ocean, they are not evenly distributed in the area between the Polar Front and the Antarctic continent. The Discovery Investigations, begun in the 1930s, outlined krill distribution on an ocean wide scale (Mackintosh 1973; Marr 1962) and results from research and fishing vessels have filled in the map of the general distribution of krill in the Southern Ocean (Dolzhenkov et al. 1988; Ichii 1990; Shimadzu 1984). The use of acoustics in recent years has provided a much more detailed picture of both local and large scale distribution and abundance of krill (Miller and Hampton 1989b).

Few estimates of the global abundance of krill have been attempted recently. This is partly attributable to a concentration on surveys of small areas of the Antarctic for either management or research purposes, and partly because of a better appreciation of the uncertainties involved in making such estimates.(Everson 1988; Everson 1992b; Everson and Miller 1994). Most authorities cite a figure of around 500 million tonnes as the krill standing stock size (Ross and Quetin 1988), although there have been indications that there may have been recent changes in overall krill abundance, at least in some areas (Siegel and Loeb 1995).

Antarctic krill, in summer, are most abundant in areas where there are physical features in the water mass or on the sea floor (Everson 1976). Krill are scarcer in the open ocean than near the ice-edge or the sub-Antarctic islands, particularly in the South Atlantic. On a large scale during summer they are found mostly in the waters of the East Wind Drift, in the shelf-break area and just to the north (Ichii 1990). In the Antarctic Circumpolar Current the krill are concentrated in the gyres formed in the South Atlantic sector and around the major island groups such as South Georgia and the South Orkneys as well as in the complex water masses around the Antarctic Peninsula (Mackintosh 1973).

The seasonal picture of Antarctic krill distribution is complicated by the lack of information available for times of year other than the summer. There is, however, a complex relationship between krill distribution and the sea ice and its seasonal movements (Daly and Macaulay 1991). Krill may follow the retreating ice-edge and its associated phytoplankton bloom during spring and remain restricted in their longitudinal distribution during summer. In winter, krill may remain where they are concentrated during summer or may disperse under the ice or retreat to the ocean floor (Gutt and Siegel 1994; Smetacek et al. 1990). Krill are found all year round in the ice-free South Georgia region and this area has been heavily fished, particularly in winter (Miller 1991).

Antarctic krill are schooling animals and swarms can cover many square kilometres and may contain millions of t of krill. Such aggregations have considerable internal structure and these aggregations may be composed of a large number of smaller, discrete schools (Watkins et al. 1990). Each school can be composed of animals of one size or sex or developmental stage and these schools can contain up to 30 000 krill m^-3 of water (Hamner et al. 1983).

Young krill are, at first, quite dependent on the currents being carried with the water mass where the eggs were laid. As they grow they become more powerful swimmers and are less at the mercy of the currents they may be able to exploit different flows of water to remain in a particular area. The extent to which this occurs is uncertain but there are distinct areas around the coast of Antarctic and the subantarctic islands where krill are perennially abundant (Mackintosh 1973). The persistence of these concentrations is likely to be a result of a complex interaction between the biology of krill and their environment.

Whether these semi-permanent large scale concentrations of krill constitute biological stocks is a matter relevant to management of the fishery (Nicol and de la Mare 1993). A biological stock is one where there is limited immigration or emigration and can generally be distinguished genetically from other concentrations of the same species. Studies which have attempted to discriminate between stocks of krill from various areas of the Southern Ocean have so far all yielded results which tend to back the theory of one, ocean-wide, stock (Fevolden and Schneppenheim 1988) but new, more sensitive genetic tests have been developed which may provide better information on krill stock identity.

4.1.2 Life history

A large number of comprehensive studies have been carried out into the biology and life history of E. superba. In particular, two monographs on the general biology of krill have been produced (Marr 1962; Miller and Hampton 1989a) and these should be consulted for an overview of the life history of Antarctic krill.

The life-history of krill has been pieced together using a combination of laboratory investigations and field studies. In late spring and during summer, female krill produce large lipid-rich ovaries. The reproductive process requires a considerable intake of food and can only take place in regions where food is highly abundant. Before spawning the weight of a ripe ovary can account for up to one third of the body weight of a female (Nicol et al. 1995). It is likely that females can produce a number of broods in one spawning season which generally lasts from late December to March (Ross and Quetin 1986). The body of the female becomes more and more distended as the ovary expands prior to spawning and after spawning a large cavity is left behind. Gravid females, rich in lipid, have been sought after by the Japanese fishery for human consumption as they are considered by consumers to have more flavour (Nicol 1989). Female krill lay up to 10 000 eggs at one time and they are fertilised as they pass out of the genital opening by sperm liberated from spermatophores which have been attached by males (Ross and Quetin 1983). Once free in the ocean the eggs begin to sink and develop. Krill eggs are extremely fragile and are rarely found in any great numbers in plankton nets so the exact vertical location and timing of spawning are difficult to pinpoint (Marschall 1983). Krill may spawn several times in a season and are likely to spawn for several seasons in their lifetime (Ross and Quetin 1986).

Krill larvae pass through a large number of developmental stages and within three to five months of hatching they begin their first winter. In autumn, the light begins to diminish, pack ice grows and further cuts out the sunlight. The water column becomes unstable and phytoplankton, the main food of all stages of krill, becomes scarce. Possible food sources for larvae in winter are ice algae, detritus and other animals of the zooplankton - they do not have the energy resources to over winter without feeding (Daly 1990).

Krill larvae emerge from the pack ice in spring more advanced than in autumn. It is probably at this point that they begin to exhibit schooling behaviour and as they develop into juveniles over their second summer this behaviour becomes more pronounced. The juveniles resemble the adults reaching some 25mm in length before their second winter starts.

Krill grow by moulting, and Euphausiids, unlike many crustaceans, moult regularly throughout their lives. In Antarctic krill, the moulting rate appears to be largely dependent on temperature though there are size and seasonal effects too (Poleck and Denys 1982). In the laboratory, at 0°C adult Antarctic krill moult approximately every 27 days whereas at 3°C they moult approximately every 14 days (Nicol 1990). Adult growth rates, measured by a variety of techniques range from -2% to +10% of body length per intermoult period dependent on season and size (Nicol et al. 1992; Quetin et al. 1994)

Krill continue to moult regularly even if starved. Laboratory studies have shown that adult Antarctic krill can survive more than 200 days of starvation and during this period they continue to moult but reduce their body size at each moult (Ikeda and Dixon 1982) The ability to survive long periods of starvation by shrinkage may be an overwintering strategy. E. superba are not particularly rich in lipids and lack significant stores of wax esters, which are often accumulated by crustaceans which experience long-term food shortages (Quetin et al. 1994). Shrinking krill may be able to survive the winter by utilising their own body protein as a fuel source. As a result, the mean size of individuals in a population at the end of winter could be considerably smaller than it was at the end of summer. Shrinking adults also begin to lose their secondary sexual characteristics (Thomas and Ikeda 1987). It thus becomes increasingly difficult to separate out size and age classes from the krill population and this hampers efforts to understand the population dynamics of this species (Nicol 1990).

The age-structure of the population is often easier to examine if the number of year classes is known and this depends on knowlegde of the longevity of krill. Early studies suggested a maximal life-span of 2 years (Mackintosh 1972) but later experimental studies extended the potential life-span to 11 years (Ikeda 1985). Other field and laboratory estimates have been reported which fall between these two extremes so although it is now thought that krill survive for considerably longer than 2 years in the wild - there is evidence for around five year classes (Siegel 1987) - it is still difficult to ascribe an accurate maximum age to krill.

Because Antarctic krill live in the open ocean and because the Southern Ocean is a difficult place to work in year round, it has not been possible to repeatedly sample the same local population with any degree of certainty over short time-scales let alone inter-annually (Quetin et al. 1994). Thus it has not proved feasible to follow changes that occur within a population of krill as it grows and over-winters in the ocean, although this has been done in embayments (McClatchie 1988). This lack of field data makes it hard to interpret laboratory-derived information on growth and moulting and to integrate these experimental results into investigations of the population dynamics of Antarctic krill.

4.1.3 The fishery

The development of the fishery for Antarctic krill has been well documented (Budzinski et al. 1985.; McElroy 1984; Miller 1991; Nicol 1989; Sahrhage 1989). Pertinent aspects of the history of the fishery are presented below together with statistics for the most recent fishing seasons. Full statistics on all Antarctic fisheries are available from the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR) Secretariat¹ in the form of Statistical Bulletins (CCAMLR 1990a; CCAMLR 1990b; CCAMLR 1996). All data on the historical catch of krill in the CCAMLR Area and its distribution and timing presented here is derived from these CCAMLR Statistical Bulletins.

¹ CCAMLR Secretariat, 23 Old Wharf, Hobart, Tasmania, 7000 Australia.

Exploratory fishing expeditions first harvested Antarctic krill in 1961/62 when two Soviet research vessels reported catches of 4 and 70 t. Throughout the 1960s ships from the Soviet Union continued to make sporadic small catches of krill as technologists attempted to develop suitable catching gear and products from krill and the fisheries scientists determined where the best concentrations lay. The catches remained small until the Soviet Union set up a permanent fishery in the Southern Ocean in 1972; 7.5 thousand tonnes of krill were landed in 1973. Full-scale commercial operations were underway by the mid 1970s (McElroy 1984).

The Japanese fishery began on an experimental basis in 1972 when a vessel from the Japan Marine Fishery Resource Research Centre caught just less than 60t in 59 days in the Indian Ocean sector. Commercial interests arrived on the scene in 1975 and catches began to rise.

Other nations have also been involved in the krill fishery but they have always been minor players compared to Japan and the Soviet Union, and later the Russian Federation and the Ukraine (Table 2).

Currently vessels from 4 nations are involved in the fishery for Antarctic krill (Table 3). The catches of krill have been unevenly distributed between Antarctic regions (Fig. 6). Total cumulative catches of krill since fishing began are approximately: 4.5 million tonnes from Area 48, 750 000t from Area 58, and 40 000t from Area 88. Currently all of the catch is being taken from the South Atlantic although a small catch (<1 000t) has been taken sporadically in Division 58.4.1 (the South East Indian Ocean sector) by a single Japanese trawler in recent years. The fishery in the South Atlantic operates in both the summer and winter (Miller 1991). Catches in the South Georgia area (48.3) reach their peak in the winter months (Everson and Goss 1991), and the fishery moves south with the retreating ice throughout area 48.2 to the southernmost fished area 48.1 in later summer (Fig. 7).

The catch of krill gradually increased during the late 1970s as the fishery moved from its experimental phase reaching a peak in 1982 when 528 201t were landed, 93% of which was taken by the Soviet Union (Fig. 8).

The krill catch underwent a major decline in the 1983-1985 season. The reason for the decline has been attributed to "technical difficulties" and it is not clear whether this was related to the discovery of the high fluoride levels in krill exoskeletons (Soevik and Breakkan 1979) or to problems in processing and marketing (Budzinski et al. 1985). A second major decline in the krill fishery accompanied the break-up of the Soviet Union. As the fisheries operations of the countries that were once part of the Soviet bloc moved onto a commercial basis, the catches of distant water, low value products such as krill began to decline. The catch by the Soviet Union in 1991 (the last year that catches were reported for this nation) was 275 495t, the combined catches of Ukraine and Russia (the only two countries of the former Soviet Union still fishing for krill) was 199 029t in 1992 and fell to 9 036t in 1993. Currently most of the krill caught in the Convention area is caught by Japanese vessels although the level of Japanese catch has been relatively stable for a number of years (Table 3). Additionally, much of the current catch by vessels from other nations also supplies the Japanese market. The low level of the fishery at present seems more dictated by lack of demand world-wide rather than by difficulties of supply in the Antarctic.

The future of the Antarctic krill fishery, in the short-term, is uncertain. Companies from countries including India, Canada, The United Kingdom, the United States, Norway and Australia have publicly expressed interest in exploiting the krill resource but there have been no significant new entrants into the fishery for over a decade. The expansion of the fishery is likely to be spurred on by advances in technology which make processing more efficient and allow full utilisation of the catch to produce a variety of products. This will have to be accompanied by marketing breakthroughs which will allow the use of krill products in a variety of industries, including food processing, aquaculture and pharmaceuticals.

4.1.4 Management

Krill fishing, and indeed harvesting of all species (other than whales and seals) in Antarctic waters is regulated by the Convention on the Conservation of Antarctic Marine Living Resources (CCAMLR). CCAMLR was ratified in 1980 and came into force in 1982 when the Commission for the Conservation of Antarctic Marine Living Resources (also referred to as CCAMLR) first met. The Convention was the culmination of two and a half year’s negotiation which began with Recommendation IX-2 of the Ninth Antarctic Treaty Consultative Meeting in 1977 that recommended that a ‘definitive regime’ for the conservation of Antarctic marine living resources should be negotiated (Edwards and Heap 1981). There are 23 full members of CCAMLR, and the Convention has been acceded to by 29 nations² including the original Antarctic Treaty signatories and most of the nations currently fishing in the area under its jurisdiction. CCAMLR's decisions, adopted by consensus in the Commission, become legally binding on Members after 180 days should no objections be lodged. France and Australia retain the right within the Convention to specify other measures as they see fit in the areas covered by 200 mile EEZs surrounding their territories in the sub-Antarctic - Kerguelen and Crozet Islands (France) and Heard and McDonald Islands (Australia) (CCAMLR 1995).

² Current Members of CCAMLR, as of January 1997, are: Argentina, Australia, Belgium, Brazil, Chile, European Economic Community, France, Germany, India, Italy, Japan, Republic of Korea, New Zealand, Norway, Poland, Russian Federation, South Africa, Spain, Sweden, Ukraine, United Kingdom, the United States and Uruguay. States which are party to the Convention but which are not members of the Commission are: Bulgaria, Canada, Finland, Greece, Netherlands and Peru.

The area within CCAMLR's jurisdiction covers approximately 32.9 million square kilometres which was intended to include the ocean between the polar front and the Antarctic continent. The subdivision of this region into Areas was conducted as part of the FAO STATLANT (statistical area) system for recording fisheries operations (Fig. 9). The further division of these Areas into subareas was mainly for the purpose of separating demersal fishing operations on banks and shelves. Because the subdivisions were made on the basis of the fisheries taking place at the time, the areas which were subdivided were mostly in the fishing areas of the South Atlantic islands and the Peninsula and the islands of the South Indian Ocean. This has resulted in the CCAMLR region being split up into subdivisions which have widely varying areas (Fig. 9).

The conflicting claims of various nations to parts of the Antarctic continent have not greatly affected the exploitation of marine living resources. Argentina and Chile declared 200 mile exclusive economic zones (EEZs) around their Antarctic claims in 1940 and Australia did the same in 1979. Adherence to the Antarctic Treaty since 1959 has prevented these nations from enforcing their EEZs. In the sub-Antarctic, north of the Treaty’s limit of 60o S France declared an EEZ around the Kergulen in 1978 and Australia followed with EEZs around Heard, McDonald Islands and Macquarie Island in 1979. The UK has recently done the same around South Georgia and the South Shetlands. The UK has used its zone around South Georgia and the South Sandwich Islands to enforce decisions made by CCAMLR, and to insist on detailed and comprehensive data reporting from these fisheries, rather than to proceed with its own management plans.

The Convention's aim was to manage resource exploitation using an "ecosystem approach" rather than to manage single species in isolation, as has been the practice in most other fisheries agreements. The adoption of this approach was largely in recognition of the profound effects that harvesting krill might have on the other elements of the ecosystem (May and Beddington 1982). The essence of the "ecosystem approach" to management can be found in Article II of the Convention which spells out the management goals in general terms:

Article II.

1. The objective of this Convention is the conservation of Antarctic marine living resources.

2. For the purposes of this Convention, the term "conservation" includes rational use.

3. Any harvesting and associated activities in the area to which this Convention applies shall be conducted in accordance with the provisions of this Convention and with the following principles of conservation:

(a) Prevention of decrease in the size of any harvested population to levels below those which ensure its stable recruitment. For this purpose its size should not be allowed to fall below a level close to that which ensures the greatest net annual increment;

(b) Maintenance of the ecological relationships between harvested, dependent and related populations of Antarctic marine living resources and the restoration of depleted populations to the levels defined in sub-paragraph (a) above; and

(c) Prevention of changes or minimisation of the risk of changes in the marine ecosystem which are not potentially reversible over two or three decades, taking into account the state of available knowledge of the direct and indirect impact of harvesting, the effect of introduction of alien species, the effects of associated activities on the marine ecosystem and the effects of environmental changes, with the aim of making possible the sustained conservation of Antarctic marine resources.

In the absence of information about biologically constituted stocks (Fevolden and Schneppenheim 1988), CCAMLR concentrated on developing management measures for the existing statistical areas. Initially, the management measures utilised information from the extensive BIOMASS (Biological Investigation on Marine Antarctic Systems and Stocks) acoustic surveys for krill carried out in the early 1980s (Everson and Miller 1994) but there has been a trend recently towards carrying out new surveys, using more advanced acoustic technology and specifically designed for the needs of CCAMLR. Acoustic surveys for krill are carried out to a standardised design using standard frequencies (usually 120 kHz, but often alongside 38 and 200 kHz) (SC-CAMLR 1991; SC-CAMLR 1992).

In the late 1980s, CCAMLR began the process of instituting precautionary management measures for the krill fishery. A precautionary limit is a catch level beyond which more complex management procedures must take over (Nicol and de la Mare 1993). It is based on the best estimates of the annual production of krill in the area and allowances have been made in the calculations for the demands of the krill-consuming vertebrates which the Convention requires be taken into account. CCAMLR has developed management measures for the krill fishery in the various subdivisions in the absence of information on the existence, or otherwise, of separate stocks in the krill population. This subdivision is a more conservative approach than developing uniform management measures for the whole CCAMLR area (Nicol and de la Mare 1993).

Precautionary limits are calculated using an initial estimate of the total biomass of the krill stock in an area, an estimate of the rate of natural mortality (including natural predation), a model of how individual krill grow in weight during their lives, and an estimate of the inter-annual variability in recruitment (de la Mare 1994a; de la Mare 1994b). This information is used to make a computer population model of a krill stock (Butterworth et al. 1991). This model has the form

Y = MB₀

where Y is the yield, M is the natural mortality rate, B₀ is an estimate of the unexploited biomass and is the proportion of the biomass that can be caught each year. The precautionary limit is calculated probablilistically using Monte Carlo simulations (de la Mare 1996). The model of the krill population, which includes random variability in recruitment, is run hundreds of times with values for growth, mortality and abundance drawn at random from suitable statistical distributions. This approach allows the incorporation of natural variability, and uncertainty in measurement. The simulation model is used to calculate a distribution of population sizes both in the absence of fishing and at various levels of fishing mortality. These distributions are used to determine the proportion () of an estimate of the unexploited biomass that can be caught each year. CCAMLR has developed a three-part decision rule for determining the value of :

1. choose ₁ so that the probability of the spawning biomass dropping below 20% of its pre-exploitation median level over a 20-year harvesting period is 10%; and

2. choose ₂> so that the median krill escapement³ in the spawning biomass over a 20 year period is 75% of the pre-exploitation median level.

³ Escapement refers to the average level of biomass of the exploited stock for a given level of fishing. Proportional escapement is the ratio of this exploited biomass to the average biomass of the stock before the start of the fishery.

3. select the lower of ₁ and ₂ as the level of for the calculation of the krill yield.

Precautionary limits utilise data that are subject to revision based on updated analyses or on new information. As a result, some of the precautionary limits set have already been revised and work carried out by the CCAMLR Working Groups will provide new information that ensures that the existing precautionary limits are constantly under review.

In 1991, at its tenth meeting, CCAMLR adopted its first Conservation Measures for krill, setting a precautionary catch limit of 1.5 million t per year for the South Atlantic sector:

CONSERVATION MEASURE 32/X

Precautionary Catch Limitation on Euphausia superba in Statistical Area 48.

The total catch of Euphausia superba in Statistical Area 48 shall be limited to 1.5 million t in any fishing season. A fishing season begins on 1 July and finishes on 30 June the following year.

This limit shall be kept under review by the Commission, taking into account the advice of the Scientific Committee.

Precautionary limits to be agreed by the Commission on the basis of advice of the Scientific Committee shall be applied to subareas or on such other bases as the Scientific Committee may advise, if the total catch in Statistical Subareas 48.1, 48.2 and 48.3 in any fishing season exceeds 620 000t.

For the purposes of implementing this Conservation Measure the catches shall be reported to the Commission on a monthly basis.

CCAMLR adopted a second precautionary catch limit for krill in 1992, this time for the South West Indian Ocean and this is reflected in Conservation Measure 45/XI which was updated in 1995 as Conservation Measure 45/XIV:

CONSERVATION MEASURE 45/XIV

Precautionary Catch Limitation on Euphausia superba in Statistical Division 58.4.2

The total Catch of Euphausia superba in Statistical Division 58.4.2 shall be limited to 450 000t in any fishing season. A fishing season begins on 1 July and finishes on 30 June the following year.⁴

⁴ The original Conservation Measure set a precautionary catch level of 390 000t, but this figure was revised following calculations carried out by the Scientific Committee using improved acoustic estimates of krill biomass. This is reflected in the higher figure in the later Conservation Measure.

This limit shall be kept under review by the Commission, taking into account the advice of the Scientific Committee.

For the purposes of implementing this Conservation Measure, the catches shall be reported to the Commission on a monthly basis.

A Conservation Measure subdividing the catch of krill in subarea 48 was adopted in 1992 (Conservation Measure 46/XI) but lapsed in 1994 when CCAMLR was unable to agree on an acceptable mechanism for subdividing the catch.

At its 1996 meeting, CCAMLR adopted a further Conservation Measure on krill covering the South East Indian Ocean sector, an area of 4.68 million km², 14.2% of the CCAMLR Area:

CONSERVATION MEASURE 106/XV

Precautionary Catch Limitation on Euphausia superba in Statistical Division 58.4.1

The total catch of Euphausia superba in Statistical Division 58.4.1 shall be limited to 775 000 t in any fishing season. A fishing season begins on 1 July and finishes on 30 June the following year.

This limit shall be kept under review by the Commission, taking into account the advice of the Scientific Committee.

For the purposes of implementing this Conservation Measure, the catches shall be reported to the Commission on a monthly basis.

Most of the areas under CCAMLR's jurisdiction which have been fished for krill (52.2% of the total CCAMLR area) are now covered by precautionary catch limits totalling 2.7 million t per year. The exception is subarea 88.1 (the South West Pacific Sector). There have, however, been questions raised about the appropriateness of continuing to use the results from the FIBEX acoustic biomass surveys in the calculation of the precautionary limits, particularly in the South Atlantic. These doubts have been raised because of methodological problems with the surveys (Everson 1992a) and because there has been evidence put forward that suggests that there may have been major changes in the krill populations in the South Atlantic over the last decade and a half (Siegel and Loeb 1995). There are also substantial mismatches between the abundance of krill estimated from acoustic surveys and that estimated from predator demand (Everson 1992b; Everson and de la Mare 1996).

A number of other general conservation measures that affect the fishery for Antarctic krill, have been passed by CCAMLR dealing with such issues as: monthly effort and biological data reporting system for trawl fisheries (Conservation Measure 52/XI) Scientific Research (Conservation Measure 64/XII), exploratory fisheries (Conservation Measure 65/XII), harvesting of stocks occurring within and outside the Convention area (Resolution 10/XII). A full list of current conservation measures is published annually (CCAMLR 1997)

Under CCAMLR there are schemes of International Scientific Observation and Inspection (see (CCAMLR 1989; CCAMLR 1993; CCAMLR 1995). The Scientific Observer scheme relies on internal or international agreements on the placement of suitably qualified observers on fishing vessels. The Inspection System allows for vessels of member countries fishing or carrying out research in CCAMLR waters, and to be boarded by certified inspectors from any of the member nations. The inspections are to ensure that activities undertaken by Members in Antarctic waters are done in accordance with the measures adopted by the Commission.

4.1.5 Ecosystem interactions

Krill have often been referred to as the pivotal species in the Antarctic ecosystem. This is because of their great abundance, and because of their central position in the food web between the phytoplankton and the large vertebrate predators such as the whales and penguins (May and Beddington 1982). The predators of krill are estimated to consume a vast tonnage of krill annually (Table 4) (data from Miller and Hampton 1989a) For comparison, the fishery reached a maximal level of just over 500 000t a year in the early 1980s.

The krill-feeding whales, historically the greatest consumers of krill, have been reduced to a fraction of their original population size and other elements of the Southern Ocean ecosystem are now more important as krill consumers. The extent of competition between the land-based vertebrates such as seals and seabirds, and the pelagic predators such as squid and whales, for the krill resource is unknown.

CCAMLR arose out of concern over the effect of an increased krill fishery on the populations of vertebrates of the Antarctic, specifically, the seals and seabirds and also the effect that it might have on the ability of the great whales to recover (Edwards and Heap 1981). The question of potential competition between the fishery and predators has therefore been a major concern to CCAMLR, especially for those predators which are restricted to land-based breeding sites during the summer (Croxall 1994). Of the many species of krill-eating birds and seals in the region, chinstrap, adelie and gentoo penguins are perhaps the most numerous, forming many densely packed breeding colonies around the continent but particularly on the Antarctic Peninsula, South Georgia and the South Shetland Islands (Cooper and Woehler 1994; Croxall et al. 1990). Typical foraging ranges for these species are 50 to 120km for adelies, 25-90km for chinstraps and 10-17km for gentoos (Trivelpiece et al. 1986). These penguin species arrive at their breeding sites in October to November, and although chicks fledge in mid to late February, chicks and adults probably continue to be restricted to areas close to breeding sites well into March (Agnew 1992). Other species such as Fur seals have much greater foraging ranges (Bengtson et al. 1991) but generally consume a lower proportion of krill in their diet. Fur seals have a slightly extended breeding period, December to early April.

Most of the investigations into the potential overlap between the krill fishery and krill dependent predators have focused on the South Atlantic where the fishery has concentrated in recent years. In an analysis of the Japanese Antarctic krill fishery, Ichii et al. (1994) have shown that the major populations of chinstraps are so distributed in the South Shetlands that the amount of direct overlap between their foraging areas and the current krill fishery off Livingston island is low. Whilst there is some overlap between fishing areas and zones of expected highest predation, even where this is greatest, the fishery currently takes much less krill than predators (Agnew 1992; Ichii et al. 1994). The precise consequences of any increased levels of fishing in currently exploited areas are dependent on the functional interaction between the predators, the krill stock and the fishery. These interactions are poorly understood at present and are the subject of continuing research within CCAMLR (Croxall 1994).

The effect of a large krill fishery on the more mobile elements of the Antarctic ecosystem - the whales and ice breeding seals - has been little studied. There is an assumption, that because of their greater mobility, they would be less affected by a localised krill fishery, however, this assumption has not been carefully examined.

4.1.6 Ecosystem monitoring

Early in its history CCAMLR realised that large-scale "global" calculations of resource availability may not be able to provide management advice on the carrying capacity of the ecosystem at small scales with sufficient sensitivity to adequately protect the resource base of the many species dependent on krill in the Antarctic ecosystem. With this concern in mind, CCAMLR considered that it should monitor the performance of critical ecosystem components in ways other than traditional stock assessment of the harvested species (Croxall 1994). This was the role chosen by CCAMLR for its Ecosystem Monitoring Program (CEMP) which was established in 1985. A number of key indicator species were chosen including: adelie penguins, chinstrap penguins, macaroni penguins, gentoo penguins, black browed albatross and Antarctic fur seal. CEMP monitors indices of predator status and breeding success, such as: penguin foraging duration, chick food ration, fledging weight, and fur seal pup growth, at a number of sites around the Antarctic (CCAMLR 1991). The program then attempts to relate changes in these indices to changes in krill availability and to distinguish between changes that result from commercial harvesting and those that are a result of natural fluctuations in the biological and physical environment (Croxall 1989). There are CEMP sites established in various areas of the Antarctic: in the Ross Sea, in the Antarctic Peninsula Region, on South Georgia and in the Indian Ocean sector (CCAMLR 1991) and data from the programs at these sites are submitted to CCAMLR each year. Attempts are now being made to incorporate CEMP data into management models (Everson and de la Mare 1996).

4.1.7 Bycatch

The question of bycatch - particularly that of larval fishes - in the Antarctic krill fishery has been a significant issue since the fishery began (Everson et al. 1991; Pakhomov and Pankratov 1994; Rembiszewski et al. 1978) particularly because of the severe depletion of some of the fish stocks in the South Atlantic. It is still uncertain whether fish bycatch is a major problem; fisheries operators reportedly avoid areas where there is likely to be a contaminating catch of fish, and large krill aggregations tend to be monospecific. There are also questions regarding the regional importance of the bycatch with some suggestions that bycatch is more significant in the Indian Ocean sector than in the South Atlantic.(Watters 1996). It has been suggested that potential problems with catching juvenile and larval fish could be avoided if the krill fishery moved off the continental shelf into more oceanic water (Everson et al. 1991; Pakhomov and Pankratov 1994). Whether this would affect the viability of the krill fishery by re-locating it in waters with less predictable aggregations of krill is unknown. The CCAMLR scheme of scientific observation has, as one of its research priorities, the estimation of the seriousness of the fish bycatch problem (CCAMLR 1993) but this requires that standardised sampling of catches and analyses of the results be carried out (Watters 1996).

4.2 North Pacific krill (Euphausia pacifica)

4.2.1 Japan
4.2.2 Canada

Euphausia pacifica is widely distributed in the North Pacific Ocean but has only been fished commercially in areas in the west of its range, off Japan, and the east of its range, off British Columbia. The Japanese fishery began by concentrating on highly visible daytime surface swarms which occurred in coastal waters. Off British Columbia, these swarms are more rare and the fishery has concentrated on sub-surface swarms of E. pacifica which occur in fjords and inlets.

4.2.1 Japan

4.2.1.1 Distribution
4.2.1.2 Life history
4.2.1.3 Fishery
4.2.1.4 Management
4.2.1.5 Ecosystem interactions

E. pacifica is fished in the Pacific region of the Japanese coastline between 36° and 40° North (Fig. 10). E. pacifica is considered a key species in Sanriku waters, the sea area off north eastern Japan, and many endemic and migrant predators including pelagic and demersal fishes, marine mammals, seabirds and benthic organisms depend on this species for food.

No systematic studies had been conducted on the general biology and the fishery for E. pacifica in Sanriku waters, until very recently, despite its important role in these waters. The North Pacific Krill Resources Research Group was organised in 1992 to conduct demographic studies, undertake biomass estimates and to develop methods for predicting fishing conditions. The members of this group include scientists from the Tohoku National Fisheries Research Institute, prefectural fisheries experimental stations at Aomori, Iwate, Miyagi, Fukushima and Ibaraki Prefectures, Tohoku University and the Ocean Research Institute of University of Tokyo. The activities of this group have greatly improved knowledge of the fishery biology of E. pacifica.

4.2.1.1 Distribution

E. pacifica occurs as far south as in Suruga Bay, 34°50’N (Sawamoto 1992) and is commercially exploited as far south as Cape Inubo, 35°40’N on the Pacific coast. The species are distributed in the whole of the Japan Sea up to the southern part of the Gulf of Tartary (Komaki and Matsue 1958; Ponomareva 1963), and in the southwestern area of the Okhotsk Sea (Ponomareva 1963, Ohtsuki 1975).

4.2.1.2 Life history

A number of studies have been carried out into the life history of E. pacifica over its wider range: off Japan (Endo 1981; Iguchi et al. 1993; Odate 1991), and off the West Coast of North America (Smiles and Pearcy 1971; Brinton 1976; Heath 1977; Ross et al. 1982). In particular, the biology of the commercially harvested population of E. pacifica off the coast of Japan has now been studied in some detail and results of these studies are summarised below.

The body length of E. pacifica landed at the Onagawa Fish Market mostly ranges from 14-19 mm (Odate 1991) with seasonal and interannual variations. Odate found a slight growth in modal length from 14.5mm in late February to 15.5mm in mid-April in 1987, but not in three other years (Fig. 11), and she suggests this may indicate that different populations visit the nearby waters of Onagawa.

Apparent growth of krill was observed during the fishing season in the coastal waters of Iwate (Taki 1993), Fukushima (Watanabe and Suzuki 1993) and Ibaraki (Ebisawa and Nemoto 1992; Yamazaki and Ebisawa 1993; Ebisawa 1994b) Prefectures. When the modal body length is plotted against the fishing season for Iwate, Miyagi, Fukushima and Ibaraki Prefectures, apparent growth seems to be limited to late April, May and June (Fig. 12). The exception is at Ryori in 1993, where apparent growth was observed during March - April.

High growth rates exceeding 0.1mm per day can be seen in late April to June. At Otsu constant growth was observed from 13 April to 13 May by sampling 5 times during this period with a growth rate of 0.13mm per day. As in 1986-1989, no constant growth was observed in the catch landed at the Onagawa Fish Market even from 1992-1994.

The growth rate of about 0.1mm per day is the maximum reported for the species off Oregon (Smiles and Pearcy 1971), southern California (Brinton 1976) and Toyama Bay facing the Japan Sea (Iguchi et al. 1993). This is surprising given the finding that swarming individuals have nearly empty stomachs (Endo 1984; Taki 1993; Taki 1994a). Fishing, however, is restricted to the daytime and examination of stomach contents from daytime samples is not sufficient to understand the overall feeding condition of E. pacifica. Good growth may be related to optimal environmental condition such as food quality and environmental temperature so simultaneous observations of these factors are needed.

In the fishing season, the sex ratio of E. pacifica tends to unity or males may slightly outnumber females (eg. Endo 1984). Brinton found apparent dominance by females in the southern California population (Brinton 1976), and he pointed out four possible reasons for this:

1. Late-maturing males of 10.5 - 11.5mm body length are without petasmas and therefore categorised as females.

2. Increasing mortality in males relative to that in females.

3. Egg-bearing females are less able to avoid capture by net.

4. Slower growth rate in females makes them increasingly numerous relative to males at successive body-length increments.

Males with empty ejaculatory duct(s) are regularly found in the commercial catch. For instance Endo (1984) found that 48% of males which were landed at the Onagawa Fish Market during the fishing season in 1978 had at least one empty ejaculatory duct with the highest value of more than 90%.

Females with spermatophores, however, are rare in the commercial catch in Sanriku waters (Endo 1984; Komaki 1967). Endo (1984) found that only 20% of females had attached spermatophores on the thelycum at the end of the fishing season in 1978, with almost no such females prior to that period.

A very high percentage of males with empty ejaculatory duct(s) combined with a low percentage of females with an attached spermatophore might be caused by unvoluntary ejaculation by males during the fishing operation, since spermatophores can easily be detached from ejaculatory ducts by handling with a needle.

The scarcity of females with attached spermatophores can be accounted for in two ways. First, copulated females leave the swarm and stay in the mid-water deeper than 15m in the daytime and therefore cannot be collected by the bow-mounted trawls originally used in the fishery (Terazaki et al. 1986). When mid-water gear was used, females were collected in abundance (Terazaki et al. 1986). The percentage of females with an attached spermatophore was as high as 45% at a station near Otsuchi sampled by simultaneous horizontal tows with nets at several depths. Similarly, 70% of females collected off Oarai, Ibaraki Prefecture, by one-boat seines had spermatophores attached (Yamazaki and Ebisawa 1993). Second, detachment of attached spermatophores from the thelycum when molting. This was supported by observations by Yamazaki and Ebisawa (1993). They found about 15% of females collected off Oarai in late March had attached spermatophores, but the carapace was not swollen. In early April, females with attached spermatophores and swollen carapaces dominated (~50%), and in mid-April, females without spermatophores, but with swollen carapaces, were most common. Therefore examination for the presence or absence of spermatophores is not enough to determine the maturity stage of females. Previous work examined only the presence or absence of spermatophore and did not pay attention to the maturity of the ovary, and therefore probably largely underestimated the abundance of mature females. The eggs and larvae of the species occur throughout the year in Sanriku waters, but most of them occur in April - July (Endo 1981.; Odate 1991; Taki 1994b).

Endo (1981) estimated the life span of E. pacifica based on seasonal samples, collected mainly with MTD nets. Sampling intervals were not short enough to estimate the life span precisely. Length-frequency histograms of females were composed of 2 peaks throughout the sampling period, and the life span was estimated to be less than three years. So females may take part in the main spawning event twice in their life. On the other hand, length-frequency histograms of males were composed of only one peak in spring, and their life span was estimated at less than 2 years. Males take part in the main spawning event only once in their life.

Odate (1991) estimated that the age of E. pacifica with a body length of 17mm, which is the average body size of landed individuals during the fishing season in Onagawa, is 1 year. She used the average growth rate of 1.0 - 1.5mm per month, which was derived from rather limited field sampling and Komaki’s data on developmental rate of early larvae (Komaki 1966).

Based on vertical haul (0 - 500m) samples collected with a Norpac net at 2 - 4 week intervals, supplemented by simultaneous horizontal tows with MTD nets, Iguchi et al. (1993) estimated the life span of the species to be less than 21 months in Toyama Bay facing the Japan Sea (Figs. 13 and 14). They found that growth did not occur in summer or winter, probably because E. pacifica could not come up to the surface to feed because of the warm surface water (> 20°C) related to the Tsushima Current in this season.

These studies all suggest a life span of more than 1 year and less than 2 years in the Japanese waters, which is similar to that in the Strait of Georgia (Heath 1977), shorter than that in the Okhotsk Sea (Ponomareva 1963) and off Kamchatka/south of Aleutians (Nemoto 1957), but longer than that off Oregon (Smiles and Pearcy 1971) and southern California (Brinton 1976).

In the coastal waters of Ibaraki Prefecture, as the surface water temperature increases at the end of the fishing season, surface swarms of E. pacifica disappear. However, swarms can be found in the cooler mid water (Suzuki 1986). Since this was discovered, mid water swarms as well as benthopelagic and surface swarms have been targeted by the fishery (Nakamura 1992).

Surface swarms appear to be formed in the years when cold water (5 - 7°C) prevails from the surface to the bottom, whereas benthopelagic swarms occur when the water column is stratified with warmer water (>10°C) at the surface and a thick, cold (6 - 9°C) deeper layer (Nakamura 1992), (Fig. 15). When the water column is stratified and the cold water layer is not thick, neither surface swarms nor benthopelagic swarms appear.

Monthly observations on benthopelagic populations of E. pacifica have been made since September 1989 at 5 stations off Nakaminato (36°20’N) (Fig. 16) (Ebisawa 1994a; Nakamura 1991). Nakamura (1991) found that the benthopelagic population appeared throughout the year on the bottom at 280m depth, with a higher catch in early spring-summer. A significant catch was collected when the thickness of the cold water layer (<9°C) was more than 55m (Fig. 17). He suggested that the fishable population comes not only from the north as previously suggested but also from deeps, namely from the benthopelagic population in this area.

Ebisawa (1994a) obtained somewhat different results: the highest catch using a beam trawl was collected in May - June in 1992 and 1993, corresponding to the late fishing period, or after the fishing period. The catch was small in February and March, before the beginning of the fishing season in the Joban - Kashima-nada area. He suggests that the fishable population of E. pacifica comes primarily from the north to this area and moves to the deeper water as the surface water temperature increases.

Recently, observations of benthopelagic E. pacifica populations by the submersibles (‘Dolphin-3K’ and ‘Shinkai 2000’ of Japan Marine Science and Technology Center) were made on the sea bottom at 200 - 300m depth off Kinkazan Island in July 1994 (Izumi et al. unpublished). The water temperature was 9-11°C at the surface and 2.5-3.0°C near the bottom. Over wide areas, dense aggregations of E. pacifica were found within 5m of the bottom. In the area where the population was most dense ophiuroids were also abundant on the sea floor, waving their ‘hands’ as if they were trying to catch E. pacifica. Ophiuroids are very abundant on the upper slope but they are rarely consumed by demersal fishes in Sanriku waters (Fujita 1992). Benthopelagic E. pacifica populations may support these high biomass, but the energy thus transferred to ophiuroids may never enter higher trophic levels.

The benthopelagic population of E. pacifica may be distributed continuously from the Joban coast to at least off Kinkazan Island. Predators such as cods, octopuses and ophiuroids may depend heavily on these benthopelagic populations. The horizontal dimension, especially the coastal-offshore range, of the benthopelagic population should be clarified to assess its biomass. Moreover, little is known on the relationship between surface swarms in the fishing season and benthopelagic population.

4.2.1.3 Fishery

Three euphausiid species are commercially exploited in Japanese waters: Euphausia pacifica Hansen, E. nana Brinton and Thysanoessa inermis (Kroyer) (Fig. 4). The catch of ‘Isada’, or E. Pacifica, is far larger than that of the other 2 species, with an average annual catch of 60 427t over the last 10 years. The average annual value of E. pacifica on landing is 1.5 - 3.6 billion yen and the fishery is now an important Japanese coastal fishery.

According to Odate's monograph on the fisheries biology of E. pacifica (Odate 1991), dip net fishing (actually using a bow mounted trawl) had been operating in Sanriku coastal waters for sand lance, Ammodytes personatus in the early Meiji Era, about 100 years ago (Photograph 1). The same method was applied to E. pacifica in mid 1940s by the fishermen of Oshika Peninsula, Miyagi Prefecture. Fishery statistics for E. pacifica have been available since 1953 at Onagawa Fish Market, but older fishermen report that the fishery began earlier.

The fishing season for sand lance and E. pacifica was from January - June, and from March - April, respectively. E. pacifica was targeted by changing the mesh of the net when the fishing conditions for sand lance were poor. The E. pacifica fishery was operated by fishermen from Onagawa and from nearby areas such as Izushima Island, Enoshima Island, Ajishima Island and Ogachi (Fig. 18).

Increasing requirements for food for sea bream culture and for bait for sport fishing in the late 1960s caused the fishery to expand to the northern and southern coasts of Miyagi Prefecture. A fishery began in Ibaraki Prefecture in 1972, and a large catch was landed in 1974 when a coastal branch of the Oyashio Current extended as far south as Inubo-zaki. A small amount of E. pacifica had been landed in the southern part of Ibaraki Prefecture earlier, but large amounts were not landed until 1974. Shortly after, Fukushima Prefecture began a krill fishery and in 1975 Iwate Prefecture became a member of the fishing prefectures. Thus the E. pacifica fishery which had begun in Miyagi Prefecture developed into an important fishery in the Sanriku (Aomori, Iwate and Miyagi Prefectures from north to south) - Joban (Fukushima and Ibaraki Prefectures) coastal waters (Fig. 10).

A small scale fishery for E. pacifica has also been conducted along the Japan Sea coast over the last 20 years off Akita, Yamagata, Shimane and Yamaguchi Prefectures and off the eastern coast of Tsushima Islands, Nagasaki Prefecture (Kuroda 1994). However, there is currently almost no catch in the above mentioned areas.

The Euphausia pacifica fishery is categorised as a Licensed Fishery which requires a licence from a prefectural governor. Small boats (less than 20t) are engaged in the fishery. One- or two-boat seines are now used in all prefectures (Photograph 2) except in Miyagi Prefecture, where originally a bow-mounted trawl was used, but since 1991 both one-boat seines and bow-mounted trawls have been used (Kuroda and Kotani 1994a).

A bow-mounted trawl can only catch swarms within 8m of the surface. On the other hand, one- and two-boat seines can catch subsurface swarms as deep as 150m by using echo sounders to detect the swarms.

The fishing period lasts generally from February to July, but varies from area to area and from year to year (Figs. 19 and 20). The main fishing period in Miyagi Prefecture is from March to April; in Ibaraki Prefecture the fishing season was similar until 1984, after that it became later than the Miyagi Prefecture - in May to June (Odate 1991; Kuroda and Kotani 1994a)

The fishing grounds are formed on the continental shelf (< 200m) within 10 - 20m from shore. The fishing depth is less than 50m in Sanriku waters, but deeper (0 - 150m) in the Joban coastal area (Kuroda and Kotani 1994a). The total annual catch of E. pacifica has increased steadily over the last 20 years (Fig. 21) exceeding 40 000t in 1978, 80 000t in 1987 and 100 000t in 1992. This increase was supported by the introduction of plastic containers in about 1975, which contain about 30kg E. pacifica, and by fish pumps in 1980’s (Kuroda and Kotani 1994a). In 1993 the total catch decreased to 60 881t, when catch regulations were imposed in Miyagi and Iwate Prefectures in order to obviate price declines. The catch was 70 575t in 1994 and 58 386t in 1995 (Ogishima 1995).

The largest catch in the 1970s was landed in Miyagi Prefecture. In the early to mid- 1980s, after the commencement of the fishery in Ibaraki Prefecture in 1972 and in Iwate Prefecture in 1975, Ibaraki Prefecture expanded its catch until more than 60% of the total catch in 1986 came from this area. After 1988, Iwate and Miyagi Prefectures were the main fishing prefectures with similar catch levels (Kuroda and Kotani 1994a; Odate 1991).

For fishermen, the most important factor related to the E. pacifica fishery is the ability to predict the length of the fishing season and the area of occurrence of the fishery. The fishing ground for E. pacifica is formed near the front between the coastal branch of the Oyashio Current and the coastal waters with optimal surface water temperature of 7-9°C (Odate 1991).

Odate (1991) found an inverse relationship between the southernmost latitude of the 5°C sea surface temperature isotherm and the annual E. pacifica catch (Fig. 22). She calculated the area of the water mass with a temperature of less than 5°C at 100m depth between 35° and 42°N and west of 145°E to be a more reliable indicator of the fishing condition for 1970-1978, and obtained a very high correlation coefficient (0.93) between the area and annual catch (Fig. 23).

Odate (1991) classified oceanographic patterns into 3 types according to the fishing conditions for E. pacifica (Fig. 24):

1. O-type (good catch years - 1977,1978,1981,1983-87). The coastal branch of the Oyashio Current flows along the Sanriku and the Joban coastal areas as south as Kashima-nada (about 36°30’N), forming a marked front against the warm water related to the Kuroshio Current. But in extremely cold years, when a water mass much colder than 5°C prevails in the Sanriku coastal area, a larger catch is landed in the Joban area than in the Sanriku area.

2. C-type (intermediate catch years - 1980,1982,1988,1989). The Oyashio Current prevails in Sanriku waters, and warm water related to the Kuroshio Current is found in the Joban area, causing a relatively narrow transition area in between the two currents. Good catches and fishing conditions are predict from the sea surface temperature in the Joban coastal area because subsurface swarms are exploited.

3. D-type (poor catch years - 1976,1979). The Oyashio Current is weak in the northern part of Sanriku waters, and the Kuroshio Current is also weak leaving a very wide transition area between.

Kotani (1992) undertook a correlation analysis between krill catches in various sea areas of the Pacific coast and the southernmost latitude of the coastal branch of the Oyashio Current. A significant correlation (p<0.01) was found between the annual catch in the whole area, the Ibaraki coast, and the Iwate coast, and the southernmost latitude of the coastal branch of the Oyashio Current. Specifically, the further southward the coastal branch of the Oyashio extends, the more catch is taken in the above mentioned areas. He also found a good correlation (p<0.01) between the southernmost latitude of the coastal branch in January and catch later in the same year. So the krill catch can be predicted by the southernmost latitude of the coastal branch of the Oyashio Current in January, about 2 months prior to the beginning of the fishing season.

Kotani (1992) assumed that the optimum temperature for E. pacifica swarming is 5 - 10°C and tried to predict the first fishing day of the year by correlation analyses. He measured the distance from the Onagawa or Otsu fishing grounds to water masses with a temperature of 10°C at 100m depth for each 10 day period from December to January and examined the relationship between this distance and the days from the mid day of each period to the first fishing day of the year. He found a significant (p<0.05) correlation between the distance from Onagawa fishing ground to the water mass and the days from mid December and January to the first fishing day of the area in the year. A highly significant correlation (p<0.01) was also found between the distance from Otsu fishing ground to the water mass and the days from late January to the first fishing date of the area in the year.

Southward shifts of the Oyashio Current seem to depend on the wind field over the North Pacific Ocean (Sekine and Suzuki 1991). Tomosada (1994) examined the relationship between atmospheric circulation indices and water temperature at several depths in the Transition area off north eastern Japan. He found that there is a negative correlation between the 4-year preceding Far East polar vortex index and the water temperature at 200m depth in the Transition area, and a positive correlation between the 4-month earlier Far East zonal index and water temperature at 50m in the Transition area. When the Far East polar vortex index is large, cold air flows down to high latitudes and sea water is cooled by the cold air. When the Far East zonal index is small, the western wind meanders and stimulates southward flow of the cooled water, resulting in a southward shift of the Oyashio Current. He showed the possibility of predicting long-term trends in the fishing condition of E. pacifica from such information.

4.2.1.4 Management

Fishery regulations are set separately for each prefecture. The license of the prefectural governor decides the fishing period, the time limit to come back to port, or operation time, fishing area, boat size and other factors (Table 5). Other regulations include total catch limit per season, and maximum number of plastic containers per boat per day. Fishermen regulate catches in order to keep the price high, collaborating with their counterparts in adjacent prefectures.

4.2.1.5 Ecosystem interactions

E. pacifica is preyed upon by almost all the commercially important fish species in Japanese waters which are endemic to, or seasonal migrant through, Sanriku waters (Endo 1981; Yamamura 1993). For instance this species is one of the main food items of Pacific cod, Gadus macrocephalus (Hashimoto 1974), walleye pollack, Theragra chalcogramma (Hayashi et al. 1968), Japanese chub mackerel, Scomber japonicus (Hatanaka 1958) and sand lance, Ammodytes personatus (Takeuchi; 1985). Fujita (1994) suggested that E. pacifica is most important as food for demersal fishes in the upper slope region, where the abundance of demersal fishes is highest but the species diversity is low, compared with continental shelf and deeper areas (> 300m) in Sanriku waters.

Kodama and Izumi (1994) compiled the data on demersal fish landings at Ishinomaki Fish Market, and examined their stomach contents to quantify the importance of E. pacifica as prey. These landings were made by an offshore trawl fishery which was mainly operated in the vicinity of Kinkazan Island with a depth range from 150-2 000m. The average yearly landing of demersal fishes at Ishinomaki Fish Market was 45 740t for the years 1983-1988. Except for 10 127t of sand lance caught in the shallower Sendai Bay area, gadid fishes dominated the catch, constituting 88% (31 355t), followed by Japanese flying squid, Todarodes pacificus 1.8% (652t), spinycheek rockfish, Sebastolobus macrochir 1.3% (460t), octopuses 0.6% (210t) and flatfishes 0.6% (203t). Gadid fishes include walleye pollack, Thelagra chalcogramma 76%, Pacific cod, Gadus macrocephalus 7%, and deep-sea rattails 17%. The main fishing depth is 100-350m for walleye pollack, 200-600m for Pacific cod and 400 - 2 000m for deep-sea rattails. Therefore, the walleye pollack is probably the most important demersal fish predator of krill in this area.

Kodama and Izumi (1994) examined the occurrence of E. pacifica in the stomachs of various size groups of walleye pollack (2 196 specimens), Pacific cod (1 089 specimens) and Pacific herring, Clupea pallasi (520 specimens) which were collected in 1973-1991. The importance of E. pacifica as food depends on the body size of the fish species. For example, the percentages of walleye pollack which included E. pacifica in their prey items were: 84% at body length of 10-19cm, 77% at 20-29cm, 60% at 30-39cm, and 42% at 40-64cm. For Pacific cod they were: 24% at 10-19cm, 47% at 20-29cm, 13% at 30-39cm, 3% at 40-59cm and 0% at 60-79cm. For Pacific herring the proportions were: 65% at 17-23cm and 70% at 24-32cm.

Pacific herring feed mainly on E. pacifica throughout their life after they have attained the juvenile stage. Walleye pollack shift their prey from E. pacifica to other species as they grow. Prey items of 40-64cm walleye pollack are: E. pacifica (42%), fishes (34%), and squids (20%). Pacific cod, on the other hand, shift from E. pacifica to other food items earlier: 30-39cm specimen feed on: E. pacifica (13%), fishes (39%), cephalopods (30%), brachyurans (22%), and other decapods (18%). Juveniles of these fish species, which feed mainly on E. pacifica, usually inhabit the 150-300m depth layers and become scarce further offshore.

Kodama and Izumi (1994) estimated the total amount of E. pacifica consumed by demersal fishes which were landed at Ishinomaki Fish Market where a large proportion of the gadid fishes caught in Joban-Kinkazan Island areas are landed. Fishery statistics are available from Ishinomaki Fish Market for several size categories of the three fish species mentioned above. Pacific herring were treated as demersal fish because they behave like a demersal fish in Sanriku waters in the southern part of its distributional range (Kodama pers. comm.).

The daily consumption of E. pacifica by these three fish species (Ck) can be given by

C_k = B_f (R_f.,/W_f) P_k

where B_f denotes fish biomass, R_f, daily ration of fish, W_f, average fish weight, and P_k, ratio of E. pacifica weight to the total weight of stomach contents. Several assumptions must be made to calculate C_k. First, fishing efficiency was assumed to be 0.3 and therefore fish biomass B_f equals annual catch divided by 0.3. Second, the daily ration was assumed to be equal to the average weight of stomach contents, since the assimilation rate is unknown. This assumption may cause an underestimation of the amount of E. pacifica consumed. Third, the ratio of E. pacifica weight to the total weight of stomach contents was assumed to be equal to the occurrence of E. pacifica divided by the sum of occurrences of all the food items.

The results are shown in Table 6. Walleye pollack consume 540 000t, Pacific cod consume 8 500t and Pacific herring 1 300t of E. pacifica per year. A total of 550 000t were consumed every year off Miyagi and Fukushima Prefectures. As 52% of demersal fish landings from the whole area from the Sanriku coast to Kashima-nada are landed at the Ishinomaki Fish Market, about 1 000 000t of E. pacifica may be consumed by demersal fishes in the whole area.

Pelagic fishes such as sardines, Japanese chub mackerel, (Scomber japonicus) and Japanese flying squid, (Todarodes pacificus) also migrate in large numbers into this area during May to January. Their biomass is estimated to be more than 1 000 000t (Kodama and Izumi 1994). Quantitative estimates of the amount of E. pacifica consumed by these pelagic fishes are also needed.

During the fishing season sand lance, black-tailed gull (Larus crassirostris) and rhinoceros auklet (Cerorhinca monocerata) feed on E. pacifica intensively (Komaki 1967). Ogi (1994) provides some information on other seabirds which feed on euphausiids in Sanriku waters. In early summer, sooty shearwater (Puffinus griseus), slender-billed shearwater (P. tenuirostris) and pale-footed shearwater (P. Carneipes) visit Sanriku waters in large flocks during their northward migration from the Southern Hemisphere. Although the abundance of rhinoceros auklet (Cerorhinca monocerata) and streaked shearwater (Calonectris leucomelas) which breed in this area, is not negligible, the sooty shearwater is by far the most numerous. Sooty shearwaters migrate into this area from late March to mid July, coinciding with the northward migration of Japanese sardine, Sardinops melanostictus and Pacific saury, Colorabis saira (Shiomi and Ogi 1992). The slender-billed shearwater is more important as a euphausiid feeder than the piscivorous sooty shearwater because the species is thought to be suited for sea surface feeding (Ogi 1994). As the slender-billed shearwater feed mainly on a euphausiid, Nyctiphanes australis, in their breeding area of Tasmanian waters (Moderhak and Cielniaszek 1991; Montague et al. 1986), it is quite possible that they consume a large amount of E. pacifica in Sanriku waters.

Although seabird observations during the winter season are rare in Sanriku waters, Ogi (1994) found large number of auklets in late February to early March in this area. These include thick-billed murre (Uria lomvia), common murre (U. Aalge), rhinoceros auklet crested auklet (Aethia cristatella) and ancient murrelet (Synthliboramphus antiquum). These auklets seem to feed almost exclusively on E. Pacifica and copepods, because there was always dense patch of zooplankton, detected by hydroacoustics, under small flocks of crested auklet. Ogi suggests that these auklets are not endemic to Sanriku waters, but came from the Okhotsk Sea to overwinter in this area. Quantitative estimates of the amount of E. pacifica consumed by these seabirds are needed.

4.2.2 Canada

4.2.2.1 Distribution
4.2.2.2 Life history
4.2.2.3 Fishery
4.2.2.4 Management
4.2.2.5 Ecosystem interactions

Surprisingly little published information is available on the fishery for E. pacifica off British Columbia despite its initiation in the early 1970s. A PhD thesis on the harvesting of euphausiids was completed in 1977 (Heath 1977) and an assessment of the fishery potential was reported in the mid 1980s (Fulton and Le Brasseur 1984). More recently, a more general account of this fishery was published (Haig-Brown 1994) but much of the information available comes from documents prepared by Fisheries and Oceans, Canada (Koke 1996). In late 1995 a workshop was held at the University of British Columbia on "Harvesting Krill: Ecological Impact, Assessment, Products and Markets". The report from this meeting has been published (Pitcher and Chuenpagdee 1995) and a book is being prepared by participants of this meeting. The workshop dealt in some detail with the euphausiid fishery off the British Columbia coast.

4.2.2.1 Distribution

About twenty species of euphausiids occur in British Columbia waters, but the krill biomass is dominated by five: Euphausia pacifica, Thysanoessa spinifera, T. inspinata, T. longipes and T. rashii. E. pacifica is typically one of the dominant species and accounts for over 70% of the euphausiid biomass in the Strait of Georgia where the commercial fishery occurs (Fulton and Le Brasseur 1984). Some commercial concentrations of euphausiids have been identified on the west coast of Vancouver Island and near the southern end of the Queen Charlotte Islands (Fulton and Le Brasseur 1984; Heath 1977). Intermittent surveys of the overall area have been accompanied by regular monitoring of krill stocks in Jervis inlet since 1990 (Fulton et al. 1982; Mackas and Fulton 1989; Romaine et al. 1996; Simard and Mackas 1989).

4.2.2.2 Life history

The main spawning season of E. pacifica off British Columbia is May to July and a second period of less intensive spawning occurs in late August-September in the Strait of Georgia (Heath 1977). The pulses in spawning activity coincide closely with the periods of high phytoplankton abundances. The life span is estimated to be 19-22 months with growth cessation occurring in autumn-winter when water temperature and phytoplankton abundances are low (Heath 1977). The estimated life span is quite similar to that obtained in the Toyama Bay, Japan, facing the Japan Sea (Iguchi et al. 1993) but longer than those off Oregon (Smiles and Pearcy 1971) and southern California (Brinton 1976). The differences in life span between the Strait of Georgia populations and their southern counterparts are attributable to the low availability of food in the Strait of Georgia during late autumn and winter and to lower water temperatures compared to the southern regions (Heath 1977). Heath recommended the period from October to March as the best time of year for E. pacifica harvesting taking into account the optimum yield and the potential impact of fishing on larvae of fish and shellfish.

4.2.2.3 Fishery

The commercial fishery for E. pacifica began about 1970 on a experimental basis and was confined to the Strait of Georgia and the east coast of Vancouver Island (Fig. 25). In 1983 participation in this fishery was restricted to those individuals who had applied for, and held, a category "Z-F" licence. This licence was not subject to limited entry. Until 1985 annual landings were less than 200t, (Figure 26) with fishing concentrated initially in Saanich Inlet, then Howe Sound and most recently in Jervis Inlet (Anon 1995). Due to continued concentration of fishing effort in Jervis Inlet rather than the adjacent waters in the Strait of Georgia, separate inlet quotas were introduced in 1989. The annual TAC increased to 785t; 500t for the Strait of Georgia and 20 to 75t for each of the major mainland inlets.

In 1990, due to concerns of local stock overfishing, the annual quota was reduced to 500t; 285t for the mainland inlets and 215t for the Strait of Georgia. In 1990, 56 licences were issued, of which 17 reported landings of 530t for a landed value of Can$415 000. This was the first year since the beginning of this fishery that the annual quota had been reached (Table 7).

Only 53t of euphausiids were reported landed in 1993 with a total landed value of Can$41 000. This decline in landings from 381t reported in 1992 was a function of market conditions rather than any decline in krill stocks. Preliminary landings of euphausiids reported for 1994 were in excess of 300t, with a value of Can$259 000 (Morrison 1995), as markets stabilised somewhat from the previous year.

The number of licences issued for this fishery increased annually from 7 in 1983 to 56 in 1990 then declined to 45 in 1991 (Table 7). In 1993 licences were limited to 25 vessels upon the advice of industry and because the annual quota was being taken by the current fleet. Only one vessel during 1993 and three vessels during 1994 reported euphausiid landings (Anon. 1994).

Two types of vessels participate in this fishery; smaller freezer vessels whose catches are limited due to freezing capacity (5-6t of krill a day) and larger vessels which land large quantities of euphausiids for onshore processing and freezing (Haig-Brown 1994). The catch must be frozen within 24hrs to avoid a significant deterioration of product quality. The fishing season can be as short as 20 days (actual fishing days) and individual vessels may land as little as 32t in a season. Nets used have mouth areas of around 80m², the trawl mouth is kept open by means of a beam and is buoyed to keep it from flipping when the ship turns and there are weights on the footline to maintain the net's shape. Fishing is carried out close to the surface - often less than 20m deep and on moonless nights when the krill rise to the surface forming layers less than 10m in vertical extent. The krill are located by echosounders. The larger vessels use a seine net and are usually out-of-season salmon fishing boats with no onboard freezing capacity. The presence of these vessels in the fishery is usually dependent on the success of the salmon fishery. If there has been a bad salmon catch, then krill are fished to increase revenues.

From 1983 to 1987 effort (hours trawling) was relatively constant and varied between approximately 300 and 600 hours annually. After 1988 there was a significant increase in effort rising to more than 1,400 hours trawled in 1991. Effort has declined since 1991 (Table 7).

Catch per unit effort (kg hr^-1) reported on harvest logs has remained relatively constant since 1986 at approximately 300 to 400 kg hr^-1. During the year of highest effort (1988) CPUE was only 255 kg hr^-1. The highest catch rates (1 153 kg hr^-1) were reported in 1992 as fishers reported fishing extremely dense concentrations of euphausiids in Malaspina Strait during the November opening. CPUE estimated for 1993 was 565 kg hr^-1 (Table 7). CPUE data do not reflect changes in efficiency through more efficient sounders and improved nets and fishing techniques. As with other species of krill, CPUE is not thought of as a particularly useful index of krill abundance and some operators have resorted to sophisticated scientific echo sounder systems to make independent stock assessments (Haig-Brown 1994).

The euphausiid fishery is market limited with the majority of the product being frozen for export to the US where it is used in the production of fish feed or pet food (Haig-Brown 1994). In 1992, as a result of change in the management of the fishery, fishers chose not to fish most of the inlet quotas early in the year and opted to take the combined annual quota from Malaspina Strait in November. Catch rates in the November fishery were exceptionally high as fishers found extremely dense swarms of krill. As a result, a large quantity of krill was landed in a short period of time and it was not possible to rapidly freeze the entire catch. Consequently, some poor quality and spoiled krill reached the market and this put off buyers. In 1993 these buyers were reluctant to purchase British Columbia krill, and landings during 1993 were low. Landings of krill during 1994 increased due to renewed market interest. It is anticipated that the fishery will proceed at the pre-1993 levels during the autumn opening. Timed openings during period of high production may be considered to ensure a high quality product.

4.2.2.4 Management

Quotas were established in 1976 for the E. pacifica fishery in response to concerns about harvesting a species upon which salmon and other commercially important finfish depend (Anon. 1977). The annual catch was set at 500t with an open season from November to March to minimise the incidental catch of larval and juvenile fish. This quota was derived from an estimate of the annual consumption of euphausiids by all predator species in the Strait of Georgia. The quota was set at 3% of this estimate. An additional 25t for Howe Sound was added to the quota in 1986 and in 1989 the annual quota was increased to 785t. This quota was subdivided: 500t for the Strait of Georgia and 20 to 75t each for several major mainland inlets. In 1990 due to concerns over declining catch rates and potential local overfishing in Jervis Inlet and Howe Sound, the inlet quotas were subtracted from the Strait of Georgia quota resulting in a quota of 215t for the Strait and a total annual quota of 500 t. In 1991, upon the advice of industry, a split season was established to allow fishing in the inlet areas during the period January to March and August 16 through December 31 and in the Strait of Georgia for the period of November through December only. This was modified in 1992, also on the advice of industry, so that the mainland inlets were open as in 1991 but to November 4 only and the Strait of Georgia opened November 4 with a quota of 215t, as in 1991, plus the balance of the unfished quotas from the mainland inlets. This was to ensure that industry would be better able to harvest the total 500t quota. Previously, portions of quotas were left unfished in some of the mainland inlets. The 1993 quotas were the same as for 1992 with the exception of Loughborough Inlet where the quota was eliminated and added to the Homfray-Price-Lewis quota. This was done in consultation with industry as fishers indicated that there were not sufficient stocks in Loughborough Inlet to support a fishery.

Quotas are managed through weekly hails of catch to fishery managers as a condition of licence. Compliance has been low and quotas in some areas have been exceeded. Prior to 1990 the total annual 500t quota had never been reached. Limited monitoring of this fishery has taken place by the Department of Fisheries and Oceans Operations Branch. The Regional Executive Committee of the Canadian Department of Fisheries and Oceans has stated that as a matter of policy the region is not prepared to support developmental fisheries on forage species such as krill. The 500t quota for the Strait of Georgia and mainland inlets will, therefore, remain fixed for the foreseeable future (Morrison 1995).

4.2.2.5 Ecosystem interactions

In 1976 quotas were established to limit the fishery due to concerns about the importance of euphausiids as a food source for salmon and other commercially valuable species. The open season was set from November to March to minimise the incidental catch of larval and juvenile fish and shellfish. Due to uncertainties of stock size and the impact of fisheries on forage species such as euphausiids, the Canadian Department of Fisheries and Oceans has limited the harvest to the Strait of Georgia and adjacent mainland inlets to a fixed quota of 500 metric tonnes. Recent studies have shown that offshore euphausiids are a major food source for offshore stocks of Pacific hake, Pacific herring, dogfish, sablefish, (Tanasichuk 1995) as well as Pacific halibut and chinook and coho salmon.

4.3 Other species of krill harvested in Japanese waters

4.3.1 Thysanoessa inermis
4.3.2 Euphausia nana

4.3.1 Thysanoessa inermis

Thysanoessa inermis has been commercially exploited since the early 1970s in the inshore waters of Shakotan Peninsula and Yagishiri Island, the western coast of Hokkaido (Hanamura et al. 1989; Kotori 1994) (Fig. 27). Surface swarms of this species are fished in the daytime usually from early March to early April. A spoon net, with a diameter of m and a 3-4m handle, is used to catch T. inermis swarms. The price varies from 75 to more than 3 000 yen per kg. The yearly catch varies from several tonnes to 200t (Figure 28). Biological characteristics of the swarming individuals were investigated by Hanamura et al. (1989) and the swarms proved to be composed of fully mature individuals. Males possessed spermatophores in the ejaculatory ducts and females spermatophores had attached at the thelycum. They concluded that T. inermis were engaged in reproductive activities during the swarming season.

4.3.2 Euphausia nana

‘Ami-ebi’ or E. nana, an allied species of E. pacifica, has been commercially exploited for 20 years in Uwajima Bay, Ehime Prefecture, Shikoku (Hirota and Kohno 1992) (Fig. 29). The distribution of this neritic species is from the southern Japanese coasts to the East China Sea and Taiwanese waters (Fig. 30). This species does not seem to be endemic to Uwajima Bay because the fishing season is restricted and the adults are reported to perform diel vertical migration with daytime depth of 300 - 400m in Sagami Bay (Hirota et al. 1983).

Fishery statistics have been available since 1976, and the yearly catch varies from 2 000 to 5 000t from 1981-1991 (Figure 31). Two fishing methods are employed to catch E. nana swarms. One is a purse seine operated at night by a troop which consists of a netting boat, a transport-boat, and up to 3 light-boats using fish gathering lamps from March to July. The other method is a boat seine operated in the daytime by a troop which consists of 2 netting boats, a boat with fish finder (with 3 frequencies of 50, 200 and 400kHz) and a transport-boat, from spring to early summer. Landed E. nana are used as feed for red sea bream. The price is about 50 yen per kg.

4.4 Krill in the North Atlantic (Meganyctiphanes norvegica, Thysanoessa raschii and Thysanoessa inermis)

4.4.1 Meganyctiphanes norvegica
4.4.2 Thysanoessa raschii
4.4.3 Thysanoessa inermis
4.4.4 North Atlantic Fisheries

These three species are often found in similar areas in the North Atlantic and fisheries have been proposed that would target all three species.

4.4.1 Meganyctiphanes norvegica

The general distribution of M. norvegica is described in (Mauchline 1967) and is extensive (Fig. 3). M. norvegica lives in water depths down to 300 m, the adults requiring water deeper than 200m whereas the younger stages may be found in shallower waters such as Passamaquoddy Bay, on the East Coast of Canada. They are thought to undergo diurnal vertical migrations throughout their range although there is evidence that during winter they may form bottom or near bottom aggregations (Greene et al. 1988). There are frequent reports of surface swarms of M. norvegica throughout its range (Mauchline 1967). The occurrence and timing of these swarms has been reported in detail in the Bay of Fundy (Bigelow 1926; Brown et al. 1979; Hollingshead and Corey 1974; Nicol 1984; Nicol 1986) in the Gulf of St. Lawrence (Berkes 1976; Berkes 1977) and in the Mediterranean, (Fisher et al. 1953). Studies on the local distribution and abundance of North Atlantic krill have used nets, acoustics and other sampling techniques (Sameoto et al. 1993). Layers of M. norvegica off the East coast of Canada have been shown to have densities up to 1 000 krill m^-3 from net sampling and acoustic studies (Sameoto et al. 1993). In the Gulf of Maine, near bottom concentrations of M. norvegica have been estimated to contain densities of greater than 1 000 krill m^-3 using acoustic techniques (Greene et al. 1988). Surface swarms have been estimated to contain higher densities, up to 770 000 krill m^-3 or up to 154 kg m^-3 (Nicol 1986). Layers of M. Norvegica can extend for tens of kilometres (Sameoto 1983) and individual surface swarms can be 30m long, can cover an area of 111m², and contain up to 2.2t of krill (Nicol 1986).

Detailed population studies of M. norvegica have been carried out in the Bay of Fundy (Kulka et al. 1982), Gulf of St. Lawrence (Berkes 1976), in the fjords of Norway (Falk-Petersen 1981), in the Baltic (Boysen and Buchholz 1984), in Scottish sea lochs (Mauchline 1960) and off Iceland (Einarsson 1945) and in the North Atlantic(Lindley 1982). M. norvegica is thought to have a 2.5 year life cycle. M. norvegica grows rapidly during its first year of life reaching a size of 25mm at adulthood. During its second year it grows more slowly to reach a final size of 40mm, some 26 months from hatching (Mauchline 1960). During winter, M. norvegica appears to enter a period of reduced growth or even size shrinkage during winter (Boysen and Buchholz 1984). The sex ratio has been found to vary markedly between catches (Hollingshead and Corey 1974) suggesting some degree of sexual segregation in the swarms in which this species lives. There is also evidence that surface swarms have a different population structure compared with the population found at depth (Nicol 1984).

Breeding occurs at different times depending on location. In the Mediterranean, M. norvegica breeds during winter and in Northern waters during summer and there may be some association between breeding and the occurrence of surface swarms in some areas (Nicol 1984). Generally, M. norvegica become mature after one year and are thought to reproduce during the following two seasons. Eggs are spawned directly into the water during a prolonged summer breeding period (Mauchline 1960) Unlike most species of krill with commercial potential, M. norvegica is largely carnivorous (Beyer 1992).

4.4.2 Thysanoessa raschii

T. raschii is usually found in water depths greater than 100m. They perform diurnal vertical migrations in the summer months and are thought to form near bottom aggregations in winter (Mauchline 1964; Zelickman et al. 1979). They have been known to form surface swarms (Zelickman 1961; Zelickman et al. 1978).

T. raschii are thought to live for over two years. In temperate waters, the eggs are spawned into the water in spring and they reach a length of 12-16mm by the winter. Maturity is reached in the second year and the animals continue growing through their second summer to reach a total length of 22-25mm by the autumn (Falk-Petersen 1981; Mauchline 1964). They may reproduce once more before dying in their third year. The seasonal cycle is somewhat altered in more northern waters, with growth being less rapid and spawning being later (Einarsson 1945).

4.4.3 Thysanoessa inermis

T. inermis is predominantly an arctic or subarctic species of krill found in shelf waters (Kulka and Corey 1978). Like T. raschii, they perform diurnal vertical migrations in the summer months and are thought to form near bottom aggregations in winter (Sameoto 1982; Zelickman et al. 1979). Surface swarming has also been reported in this species in the North Atlantic (Stott 1936) and in the North Pacific (Hanamura et al. 1989).

T. inermis appears to live for 2 years through most of its range and spawns in summertime at the end of its first and second year. They grow to a size of approximately 18mm after their first year of growth and reach a size of 22mm after their second year (Kulka and Corey 1978; Kulka 1978; Falk-Petersen 1981).

4.4.4 North Atlantic Fisheries

4.4.4.1 Gulf of St. Lawrence
4.4.4.2 Scotian Shelf

4.4.4.1 Gulf of St. Lawrence

The Gulf of St. Lawrence is the only area currently fished for krill in the North Atlantic (Figures 3 and 5). The potential for harvesting krill in the Gulf of St. Lawrence was first examined following the depletion of traditional fish stocks in the area and a program was started in 1972 to locate large concentrations of krill in the Gulf (Runge and Simard 1990). Three species are commonly found in the Gulf of St. Lawrence: Meganyctiphanes norvegica, Thysanoessa raschii and Thysanoessa inermis (Simard et al. 1986b). The estimated biomass of krill in two areas of the Gulf where krill are most concentrated was 75 000t (Sameoto 1975) and an estimated catch rate for trawlers fishing a 100m² mouth opening trawl was estimated to be 379 kg h^-1 based on a biomass estimate of 1g m^-3. The estimated potential for exploitation of all three krill species in the Gulf, based on an exploitation rate of 50% of the biomass, was 37 500t estimated in 1975 to be worth Can$3.75 million (Sameoto 1975).

More recent studies have used acoustics to determine the abundance of krill in the Gulf of St. Lawrence (Simard 1995; Simard et al. 1986a; Simard et al. 1986b) and acoustic estimates of krill biomass for krill in the Gulf range from 400 000t to 1 million tonnes.

In 1991 the Canadian Department of Fisheries and Oceans (DFO) issued a scientific permit to fish zooplankton to study the quality of zooplankton in the Gulf of St. Lawrence (Runge and Joly 1995). An exploratory fishery for krill and the copepod Calanus finmarchicus subsequently took place in the Laurentian Channel. The permit was renewed in 1994 with a "preventative" TAC of 100t for krill and 50t for Calanus. The fishery took place in November when 6.3t of krill and 400kg of Calanus were harvested (Runge and Joly 1995).

The Gulf of St. Lawrence krill fishery is managed by DFO, Laurentian Region. The fishery in this region is still experimental or pre-commercial (Runge and Joly 1995). In order to undertake a pre-commercial fishery, a fisherman must show DFO that he is an experimental fisherman, that he knows how to preserve the catch to avoid loss and that he can sell them at a acceptable price. No undue loss of catches is tolerated either at sea or when stored. During krill fishing there is an observer on board who reports on every detail of the catch including bycatch. The krill fishery in the Gulf is based on the DFO policy for underutilized species. The management of the Gulf krill fishery is based on a best guess of the biology of the harvested species with a bias towards a conservative approach in each case. Permits to enter the krill fishery are issued based on policy guidelines developed by DFO Eastern Canada regions (Gendron 1994) and these include:

· preparation of a strong management plan which reflects a common position taken by both management and science.

· interaction with the industry so that management objectives can be clearly understood by all parties.

· acquisition of some basic biological data by developing partnership with selected fishermen.

· formal financial commitment should be made over a 3-4 year period to biologists concerned with the potential fishery for development work.

· information on production, long-term sustainability of the fishery, distribution of the resource, stock discrimination, patterns of aggregation.

· experimental management being recognised as one of the best forms of large-scale experiment.

Biomass estimates for krill in the Gulf range from 400 000t to one million tonnes and the TAC of 300t is based on these biomass estimates and reflects the uncertainties involved. It has been assumed that the catch level - of the order of one half of one percent of the minimum estimate of the available biomass - would have a negligible effect on the krill populations and on the populations of natural predators on krill. Factors that have been taken into account when designing management strategies for krill in this region include: the problems of taking the whole of the catch from a restricted area, the effect on the populations of whales that feed in the area and the incidental bycatch, particularly of juvenile fish (Runge and Joly 1995).

The Gulf fishery produces frozen krill and freeze dried krill for ornamental fishes and for public aquaria and freeze dried krill as an ingredient in salmon feed and as a flavourant for food for human consumption.

Krill in the Gulf of St. Lawrence are fed upon by a variety of fish: capelin, herring, sand lance, mackerel, cod, redfish and whales for flatfishes. They are also the prey of whales - particularly blue, fin, minke and humpback and a variety of seabirds (Simard 1995). There is particular concern that any development of the krill fishery in this area should not interfere with the local whale-watching industry (Y. Simard, pers. comm.).

4.4.4.2 Scotian Shelf

A permit to fish 1 000t of krill (primarily M. norvegica) on the Scotian Shelf and Gulf of Maine, off Nova Scotia, Canada, was requested in 1995. The krill would be used to produce a product to coat fish pellets to be fed to young salmon in fish farming (Harding 1996). On the Scotian Shelf, there are additional concerns about the effect of the proposed krill fishery on the fish species of the region which have a major portion of krill in their diet. There is also considerable concern about the possibility of a significant by-catch of larval and juvenile forms of other commercial species that could be taken with the krill catch and possible interactions with populations of the endangered right whale is of utmost importance in regard to decisions about allowing a krill fishery off Nova Scotia.(Harding 1996).

4.5 Other potential krill fisheries

4.5.1 Nyctiphanes australis in Australian waters

4.5.1 Nyctiphanes australis in Australian waters

The commercial potential of Nyctiphanes australis in Tasmanian waters, particularly as an aquaculture feed has recently been appraised (Virtue et al. 1995; Virtue et al. 1996).

N. australis is distributed on the continental shelf of South East Australia and on the New Zealand continental shelf (Fig. 2) (Blackburn 1980). It is one of the more inshore species of krill and is frequently found in swarms in coastal waters or washed up on beaches (O'Brien et al. 1986). N. australis is generally found in waters between 200m and the surface but has sometimes been reported from depths as deep as 400m. It is typically found in sub-surface aggregations but surface swarms and benthic aggregations are not rare, particularly in Tasmanian waters (O'Brien 1988). There is some evidence of diurnal vertical migrations by this species but generally it is found in aggregations throughout the upper water column both by night and day (Young et al. 1993).

N. australis is one of the smaller species of krill reaching a maximum size of only 20mm and a maximum weight of 40mg after about one year (Ritz and Hosie 1982b) It is also one of the species of krill that broods its eggs until they hatch rather than spawning them directly into the water column. N. australis reach sexual maturity after about 4 months and the females may lay several broods of eggs in one season (Hosie and Ritz 1983).

N. australis occurs in dense aggregations close inshore off the coast of Tasmania. There are estimated to be large stocks of N. australis in this area (Ritz and Hosie 1982a) with figures for annual production of around 100 000t/yr (Virtue et al. 1995). The fishery for jack mackerel has been one of Australia's biggest in recent years and much of this catch is directed to fishmeal for aquaculture which is an expanding industry in Tasmania (Williams et al. 1987) Jack mackerel are almost wholly dependent on N. australis for food (Young et al. 1993) so any fishery-induced reduction in N. australis stocks has the potential to affect the jack mackerel catch. Given that both fisheries would be providing product to the same market, a fishery at the lower trophic level might be expected to yield a higher catch.

N. australis is one of the most important dietary items for a number of vertebrate species in Tasmanian waters (Ritz and Hosie 1982a). These include: jack mackerel (Trachurus declevis) where it constitutes 99% of the diet (Webb 1976; Young et al. 1993), short tailed shearwater (Puffinus tenuirrostrus) fairy prion (Pachyptila turtur), Australian salmon (Arripestrutta), skipjack tuna (Katsuwonis pelamis), tiger flathead (Platycephalus richardsoni), barracouta (Leonuro atun) and the slender tuna (Allothunnus fallai) as well as other abundant fishes and seabirds. Decisions on whether to prosecute this fishery should consider the implications for other fisheries in the area as well as the impact a krill fishery might have on the bird populations in the region.