Deepwater fish resources in the northeast Atlantic: fisheries, state of knowledge on biology and ecology and recent developments in stock assessment and management

P.A. Large^[42] and O.A. Bergstad^[43]

1. INTRODUCTION

Most deepwater fishes are long-lived, slow-growing, have a low reproductive capacity and are adapted to live in an ecosystem of low energy turnover in which major environmental changes occur infrequently. Deepwater fishery resources are, therefore, highly vulnerable to exploitation (Merrett and Haedrich 1997, Koslow et al. 2000, Anon. 2001) and deepwater habitats are sensitive and in need of protection (OSPAR 2000). Experience in the South Pacific and elsewhere has shown that deepwater fish stocks can be depleted quickly (Koslow et al. 2000) and that recovery can be slow (Anon. 2001). In most cases, reliable information on stock status and fisheries production potential has lagged considerably behind exploitation (Large et al. 2003). In the Northeast Atlantic this concern has been exacerbated by the fact that until 2003 most fisheries were completely unregulated.

Deepwater species in the Northeast Atlantic are not always clearly identified in ICES (International Council for the Exploration of the Sea) literature. This is largely because of operational decisions regarding the structure and development of assessment Working Groups (Gordon et al. 2003) and because some species are confined to deepwaters, defined as depths greater than 400 m, while others are also found in shallower waters on the continental shelf. The deepwater species referred to in this paper are those under the remit of the ICES Working Group on the Biology and Assessment of Deep-sea Fish Resources (WGDEEP) and include examples of the former, e.g. orange roughy (Hoplostethus atlanticus) and roundnose grenadier (Coryphaenoides rupestris), and of the latter, e.g. ling (Molva molva) and tusk (Brosme brosme).

Formalized collation and examination of fisheries and biological data of Northeast Atlantic deepwater species for assessment purposes began in 1994 when the ICES Study Group on the Biology and Assessment of Deep-Sea Fisheries Resources was first convened (Gordon 1998). In 2000, the group was re-established as an ICES Working Group. Since 1994, the Group has reported annually to the ICES Advisory Committee on Fishery Management (ACFM). Much of the information presented in this paper has been derived from information contained in the Study/Working Group’s reports.

2. OVERVIEW OF DEEPWATER FISHERIES IN THE NORTHEAST ATLANTIC

Comprehensive overviews of deepwater fisheries of the Northeast Atlantic are compiled every second year in the reports of WGDEEP. The reports also include tables of landings by species and ICES Sub-areas and Divisions. Recent reviews of historical trends were also provided by Gordon (2001) and Gordon et al. (2003). Hence only a brief account of historical trends and recent developments will be presented here.

Most current deepwater fisheries originated as artisanal fisheries off Portugal, southern Spain and the Azores, but also in the deep shelf areas off northern European countries, especially Iceland and the Faroe Islands. Many such fisheries using traditional gears (handline, longline) still exist, but most deepwater fish landings today are from highly mechanized longline or trawl fisheries. There has been a steady enhancement of vessel and gear technology and dedicated exploration of new grounds, often subsidised by national governments. The major expansion and industrialisation of deepwater fisheries started after World War II.

The present mechanized longline technology was mainly developed in the Nordic countries and longlining for ling, tusk and halibut (Hippoglossus hippoglossus) in deepwater has a history back to the 1860s (Bergstad and Hareide 1996). New grounds were explored in phases, and the modern longliner is essentially a factory ship of 100 feet or more long, equipped with automatic baiting systems (autoline) and storage facilities, permitting 6-8 week trips or more. Norway alone operates 50-60 such vessels, and the same technology has been adopted by many other nations, including the Faroe Islands, Iceland, Russia, Spain and Ireland.

Trawl fisheries developed mainly in the North Sea but, compared with longline fisheries, the expansion into deepwater occurred later, initially after major USSR exploratory effort in the 1960s and 1970s (Pechenik and Troyanovsky 1970, Troyanovsky and Lisovsky 1995). Western European fleets were encouraged to move into deeper water by the loss of opportunities in traditional shelf fisheries and it was French trawlers that started major commercial deepwater operations in the mid-1980s, fishing for blue ling (Molva dypterygia), roundnose grenadier and, later, orange roughy. German fleets concentrated mainly on redfish (Sebastes spp.) and the UK developed a longline fishery for deepwater sharks in addition to taking a bycatch of deepwater species in fisheries for anglerfish (Lophius spp.) and hake (Merluccius merluccius). Current trawl fisheries are mainly carried out by French, Spanish, Faroese, Irish and Scottish vessels. Russian and Polish vessels are also active, especially on the Mid-Atlantic Ridge and on slopes of the Rockall and Hatton Banks to the west of Scotland and Ireland. Deepwater trawl fisheries have not to any great extent been conducted in the southern parts of the ICES area, such as off Spain, Portugal, the Azores and Madeira, where longline fisheries persist.

Most artisanal and industrialized deepwater fisheries target benthopelagic species, primarily aggregating species or species that are easily attracted to bait. However, there are also fisheries with semi-pelagic or pelagic trawls for mesopelagic species such as blue whiting (Micromesistius poutassou) and greater silver smelt (Argentina silus). Some typical seamount fisheries for alfonsino (Beryx spp.) are also semi-pelagic in nature. Many of the demersal trawl and longline fisheries are mixed fisheries and have large bycatches of non-target species.

It is currently impossible to provide a full historical overview of fleets, effort and landings of all the deepwater fisheries of the ICES area. In 2003, NEAFC initiated compilation of such data based on national statistics. Landings statistics compiled by ICES Sub-areas and Divisions (Figure 1) are considered to be relatively complete, but a simplified overview for the period 1988-2002 is presented in Figure 2. The data from 2001 and 2002 are preliminary and may be incomplete.

There are relatively stable geographical patterns in the species composition of landings. In the northern sub-areas, ling and tusk remain the main target species and dominate the landings, and these two species are also prominent at Iceland and around the Faroe Islands. Overall, these species, plus blue ling, constitute about half of the total landings of deepwater species from the ICES area. West of the British Isles (Sub-areas VI, VII), roundnose grenadier, blue ling and orange roughy are prominent and to the south of this, black scabbardfish (Aphanopus carbo) and sharks are important. On the Mid-Atlantic Ridge (mainly Sub-area X), alfonsino is mainly fished in the south and roundnose grenadier in the north.

The major expansion of deepwater trawl fisheries was in the mid-1980s in Sub-areas V and VI and subsequently in Sub-area VII. Aggregations of blue ling were important initial targets and in many areas such aggregations were soon heavily exploited or depleted. Orange roughy fisheries started in the early 1990s in Sub-area VI, but after a few years landings dropped significantly. During the past few years orange roughy has mainly been taken in Sub-area VII, and the fisheries there are extensive but probably unsustainable. Another recent development is the expansion of fisheries on the most western bank off Europe, the Hatton Bank. On the Mid-Atlantic Ridge, alfonsino stocks around seamounts in international waters appear to be in a poor state (Anon. 2002a), and roundnose grenadier fisheries continue at low levels. Off the southern European shelf, traditional longline fisheries for black scabbardfish and sharks appear relatively stable and there are also new fisheries for blackspot seabream (Pagellus bogaraveo) off southern Spain. However, several decades ago the latter species was the target of Spanish and French fishery, producing annual landings of several tens of thousand tonnes. The resource was probably overfished and has never recovered.

3. CURRENT STATE OF KNOWLEDGE ON FISH BIOLOGY AND ECOLOGY

Knowledge of the biology and ecology of the target deepwater resources and communities remains limited, and the peer-reviewed literature is relatively small. Major international projects funded by the European Union or Nordic Council focusing on biology and ecology were conducted in the 1990s (e.g. Magnússon et al. 1997, Anon. 2000, Gordon 2001, Menezes et al. 2001), but research activity has since declined and now depends heavily on scarce national funding. ICES WGDEEP has compiled available information on growth and reproduction for most target species and tables and references are available in the reports and are not repeated here. The deepwater fish currently exploited range from extreme K-strategists, such as orange roughy and deepwater sharks, to relatively fast-growing and fecund species such as ling, alfonsino and blue whiting. Based on available information on life history strategies, it is possible to rank most species on an r-K scale, but the problem remains that too much of the information on basic biology is documented only in the grey literature. Further, data were too often collected from limited geographical areas compared with the known ranges of the species or communities (and even the known fishing areas), and there are many questions about their representivity. For some species there are major gaps in our knowledge of their basic biology. Age determination problems persist for many species and validation of age is required for most species.

In summary, priority areas for future research of immediate interest to stock assessment and management remain

• age determination and validation

• improved documentation and geographical distribution of information on growth, reproduction and mortality, and

• improved knowledge on past and present distribution (spawning and nursery areas, egg and larval drift, aggregation areas), and the identity and structure of populations and stocks (population genetics).

FIGURE 1. ICES sub-areas and divisions

Many research projects have provided knowledge but, owing to limited duration and resources, a number of these studies were discontinued too early, in essence when they had reached a stage of maturity at which breakthroughs could have been made. There are major challenges for the future, and if current competence and new technologies and methods are adopted, it may be possible to fill knowledge gaps and discover much more.

Some recently published studies include investigations of behaviour and abundance using direct observations from submersibles (e.g. Uiblein et al. 2003) and the application of otolith microchemistry analysis to study population structure (e.g. Swan et al. 2003). New initiatives such as the MAR-ECO project on the Mid-Atlantic Ridge has elements studying life history biology, ecology, and population genetics of fishery resources and also include comparisons with results from continental slope waters (Bergstad and Godø 2002). This, and other projects, may help revive deepwater fish research in the North Atlantic.

4. RECENT DEVELOPMENTS IN STOCK ASSESSMENT

Stock assessments of the major commercially exploited deepwater species of the Northeast Atlantic were first attempted at ICES in 1998 (Anon. 1998), and the most recent assessments were carried out by WGDEEP in 2002 (Anon. 2002a). While progress has been made across this period, many of the problems experienced in earlier years still persist.

As described above, little is known about the stock structure of most species and for assessment purposes, stock units have been defined on the basis of current knowledge of species distribution and similarity of catch-rate trends among ICES statistical areas (Anon. 1998). Therefore, current stock units comprise individual or groups of ICES Sub-areas and occasionally ICES Divisions. This is far from ideal because ICES statistical areas were devised for the continental shelf and, in some instances, are completely inappropriate for delineating deepwater stocks. One solution would be for countries to report catch data by ICES statistical rectangle, allowing assessment scientists to aggregate data by whatever stock areas they deem appropriate. However, rectangle data may not be available historically for some series, and some countries may have confidentiality concerns related to the release of information from new and developing fisheries. A second option is to retain the existing ICES Sub-areas and Divisions, but to subdivide them into groups of rectangles on the basis of topographical features, depth and the spatial distribution of stocks. ICES and the Northeast Atlantic Fisheries Commission (NEAFC) are currently exploring these options.

A further problem is that there are few fishery-independent surveys designed to provide time-series abundance data for use in assessments. Experience with assessments of orange roughy stocks off New Zealand and Australia has shown that assessments are likely to be more robust if a range of fisheries independent data are available, from acoustic, trawl and egg production surveys for example. Most deepwater surveys in the Northeast Atlantic have either been exploratory in nature or designed to collect biological data, and consequently assessments rely on abundance indices derived from commercial catch and effort data. These data are often sparse, are sometimes of poor quality and are not always available to WGDEEP.

No progress was made with assessments of ling and tusk in 2002 because effort and catch-per-unit effort (CPUE) series for the Norwegian and Faroese fisheries could not be updated because of lack of reporting (Anon. 2002a). Problems were also encountered in assessments of deepwater stocks to the west of the UK, nearly all of which rely on abundance indices derived from catch/effort data from the French deepwater trawl fleet, the dominant fleet in these fisheries. Prior to 1999, these indices, which are of reasonably good quality and date back to the early 1990s, were calculated using fishing effort directed specifically at deepwater species. However, in 1999 the French national database was reformatted and the data could not be accessed. As an interim measure, revised time-series data based on total rather than directed effort were presented at WGDEEP in 2002. While there was some agreement on historical trends for most stocks between the old and new series, estimates of annual CPUE for 2001 were extremely high for both black scabbardfish in Sub-area VII and deepwater sharks (mainly Portuguese dogfish Centroscymnus coelolepis and leafscale gulper shark Centrophorus squamosus) in Sub-area VI. These values indicate a dramatic recovery of these stocks in a single year, almost back to pre-exploitation levels but this was considered to be unlikely given current understanding of their stock dynamics. WGDEEP will review this position when it meets in 2004. It is anticipated that the French time-series based on directed CPUE, fully updated to include values from 1999 onwards, will be available to the Group then.

The methods used in assessments are largely determined by the availability and characteristics of fisheries and biological data. In addition to CPUE data from commercial fishing vessels, for most stocks the only other time-series data available are of total international landings. Options for assessment methods are therefore somewhat limited and the main method used has been depletion modelling using surplus production and modified DeLury models (Anon. 1998, 2002a). These models provide estimates of current and virgin exploitable biomass from which a ‘depletion ratio’ can be calculated for each stock. A full description of the strengths and weaknesses of the methods, particularly in relation to deepwater stocks in the Northeast Atlantic, is provided by Large et al. (2003). Varying degrees of success have been experienced, largely depending on the degree of contrast in abundance data and the length of time-series data available. Frequently the fit of the models has been poor, resulting in population estimates with wide confidence limits. Under these circumstances ICES advice on the state of stocks has been based on depletion ratios calculated by use of a smoothed time-series of CPUE from commercial fishing vessels as an index of exploitable biomass. These data frequently show a strong and persistent decline, and the latest ICES scientific advice is that most stocks are harvested outside safe biological limits (Anon. 2002b).

It is recognized that there is a need to expand the range of methods used in assessments to include methods used on deepwater stocks in other parts of the world (Large et al. 2003). The use of stock reduction models, which have been applied widely in assessments of deepwater stocks off New Zealand and Australia, is currently under investigation (Large 2002). These methods use biologically meaningful parameters and information for time delays attributable to growth and recruitment to predict the basic biomass dynamics of age-structured populations without requiring information on age structure. Therefore, they can be considered to be a conceptual hybrid between dynamic surplus production and full age-based models (Hilborn and Walters 1992). The model that has been trialed is part of program suite (PMOD) developed by Francis (1992, 1993) and Francis et al. (1995). Investigations have largely been restricted to studies of orange roughy in ICES Sub-area VI and results have been similar to those obtained from surplus production and modified DeLury models. Current exploitable biomass is estimated to be around 25 percent of virgin biomass and this is consistent with data from the fishery. The model also provides an estimate of the annual mean catch that can be taken, consistent with a 10 percent probability of spawning stock biomass (SSB) falling below 20 percent of virgin SSB. In New Zealand and Australian fisheries this catch is termed the maximum constant yield (MCY). Estimates of MCY for orange roughy in Sub-area VI indicate that sustainable catches may be as low as 1.5 percent of virgin biomass, probably <100 t per annum. This contrasts with annual catches taken in the early years of the fishery of up to 3500 t. Stock reduction models are being tested on other stocks in preparation for the WGDEEP meeting in 2004.

Opportunities to use a wider range of assessment methods, including length- and age-based methods, are limited. Although some progress has been made in length and age sampling of commercial deepwater landings, time-series data are available for few species and are often too short (<5 years) or incomplete (missing years) to use in assessments. Moreover, length composition data rarely exhibit any multi-modal structure or evidence of modal progression and length-based assessment methods, which rely on a strong link between modal length and age structure, are therefore unlikely to be suitable. It is possible that time-series catch-at-age data for ling at the Faeroes (ICES Division Vb) and red (blackspot) seabream at the Azores may soon be sufficient to attempt virtual population analyses using abundance indices from Faroese longliners and an Azorean longline survey, respectively. Further, if problems with age determination of blue ling can be overcome and the biological sampling of roundnose grenadier continues, age-based assessments of these species may be possible in future (Large et al. 2003).

The only information currently available on the level of fishing mortality (F) in deepwater fisheries is from estimates of total instantaneous mortality rates (Z) calculated from catch curves fitted to annual length or age compositions of species for which estimates of natural mortality (M) are available. Such analyses assume that F, M, catchability and recruitment have remained constant over time and only provide information on the scale of F, whether it is high or low, rather than accurate estimates. Critically, the depletion and stock reduction models currently used in assessments do not provide any information on the relationship between F and catches. This situation has important implications for fisheries management.

5. RECENT DEVELOPMENTS IN FISHERIES MANAGEMENT AND THEIR IMPLICATIONS

Almost all deepwater fisheries in the Northeast Atlantic were, until 2003, unregulated. Exceptions included fisheries where individual countries introduced national management measures for specific species, licensing of vessels fishing for greater silver smelt in Norwegian waters for example, and where for historical reasons individual species have been assessed and managed internationally, Greenland halibut (Reinhardtius hipploglossoides) and redfish for example. Most other fisheries, and particularly those on the high seas, followed or, on the basis of available data, appeared to be following, a boom-and-bust trajectory. Typically, exploratory fishing identified new fisheries and these were fished down outside safe biological limits over varying time-scales: short - < 5 years, for species associated with seamounts and other topographical features, orange roughy for example, and longer, perhaps 20-30 years, for more widely distributed species such as blue ling. Again, there are exceptions to these patterns. For example, the long-standing, artisanal, longline fisheries for black scabbardfish off the Portuguese mainland and Madeira is considered, on the basis of available evidence, to be sustainable. In overall terms, however, fisheries management bodies in the Northeast Atlantic, as in many other parts of the world, have been slow to follow the Precautionary Approach (Garcia 1994) and bring in management measures for deepwater fisheries. In this paper we will not review the reasons for this but will concentrate on the content and likely efficacy of management measures recently introduced.

In January 2003, the EU introduced Total Allowable Catches (TACs) for some species (EC 2002a), and, as a first step towards effort management, a vessel licensing scheme with aggregate power and capacity of deepwater fishing vessels capped to levels observed in the years 1998-2000 (EC 2002b). These measures apply to EU vessels fishing in EU and international waters. In addition, NEAFC has introduced a temporary freeze on effort in deepwater fisheries in the Regulatory Area (Figure 3) from 1 January 2003 and is currently developing further management measures.

The attractions of using TACs as a fisheries management instrument are well documented (Cochrane 2002). Most fisheries are shared internationally and managers and politicians find TACs attractive because they provide a simple mechanism for the allocation of national quotas. Moreover, establishing track records for national quotas is more straightforward because historical landings data are often more readily available than effort data. Notwithstanding this, managing deepwater fisheries in the Northeast Atlantic solely by TACs and quotas is unlikely to be successful because the relationship between catches and fishing mortality is not known, and a reduction in TACs may not result in a commensurate reduction in fishing mortality. The TACs recently introduced by the EU, although for some species significantly lower than recent total international landings, may not have a strongly positive conservation effect because most stocks are declining and catches may drop because of falling catch rates rather than because of reductions in fishing mortality. Also many commercially important exploited species, such as deepwater sharks, alfonsino and greater forkbeard (Phycis blennoides), are not covered in the TAC Regulation.

A further concern is that TACs will lead to more discarding in mixed fisheries. In many deepwater fisheries, catches consist of a range of deepwater species or, as on the continental slope west of the UK, a combination of deepwater and continental shelf species such as hake (Merluccius merluccius) and anglerfish (Lophius spp.). In these fisheries, TACs for different species can be taken at different rates, leading to increased discards of over-quota species by vessels fishing on under-subscribed species. This will lead to increased fishing mortality on deepwater species because all discarded fish die as a result of changes in pressure as they are brought to the sea surface and also, in trawl fisheries, because of abrasion in nets (most deepwater species lack a mucus covering and are susceptible to damage by abrasion). Non-commercially important species taken as bycatch will also die, and so fishing, and trawling in particular, can have a considerable impact on the wider deepwater fish assemblage.

Managing deepwater fisheries in the Northeast Atlantic by effort control also has advantages and disadvantages. The main advantage is that reductions in fishing effort would have a proportionate effect on fishing mortality, and allow management of fisheries in the absence of any knowledge of the relationship between fishing mortality and catches. Regulation of species-specific effort may be appropriate for directed fisheries, but in mixed fisheries, fleet- or gear-specific measures may be required (Anon. 2003). An important disadvantage of effort control, however, is that in a restrictive effort regime, fishers may respond by increasing the efficiency of fishing and fishing gear, a process known as ‘technological creep’. Improvements in gear design, fish-finding and navigation equipment can be important drivers of improvements in fishing efficiency in coastal artisanal deepwater fisheries, whereas in high-seas fisheries the development of fishing gear and fishing techniques is ongoing. Modest increases of between 2 and 4 percent a year will lead to a doubling in fishing efficiency in about 36 and 14 years, respectively. The effects of technological creep can be minimised by monitoring vessel and gear characteristics and limiting change, but this approach stifles innovation, requires extensive data collection, and is unlikely to be fully effective because fishers will always try to find some way of ‘getting an edge’ over other fishers and particularly over management regulations (Pope 2002).

Fisheries management bodies in the Northeast Atlantic recognize that technical measures such as mesh size regulation and selectivity grids are unlikely to be effective for deepwater fisheries because of the unusual shape and size of some species and also because a high proportion of fish entering trawls and subsequently escaping through meshes will be subject to abrasion and die (Connolly and Kelly 1996, Koslow et al. 2000). They are also aware that the use of closed areas may be appropriate for protecting spawning concentrations, of blue ling for example, and that no-trawl zones may protect fish species and coral communities associated with seamounts. However, these features can be widely scattered geographically and, although satellite monitoring would help, a patchwork of closed areas may be extremely difficult to effectively enforce. ICES is currently collating available information on these features (Anon. 2003), but data from dedicated scientific surveys are sparse. Moreover, high-resolution data from commercial vessels is difficult to obtain because of confidentiality concerns of skippers and vessel owners, particularly in new and developing fisheries.

Therefore, the best way to manage these fisheries may be by a combination of fishing effort and catch controls and closed areas/no-trawl zones. Compared with fisheries on the continental shelf, more emphasis should be placed on fishing effort, given the restricted nature of results from stock assessments and the problems with managing mixed fisheries by TACs. However, factors affecting fishing efficiency will have to be closely monitored. The management measures introduced by the EU and NEAFC are, therefore, a step in the right direction. However, they are not sufficiently rigorous in their scope and content to reduce exploitation to sustainable levels. Most deepwater stocks were identified as outside safe biological limits in 1998, so capping effort at levels observed between 1998 and 2000 will do little to conserve stocks. Further, capping effort at current levels, as applied by NEAFC, will do even less. The boom-and-bust cycle, seen in many fisheries, will simply continue.

A major weakness of the management introduced measures is that little provision has been made to control and regulate new and developing fisheries. Current ICES advice is that ‘fisheries be permitted only when they are accompanied by programs to collect data and expand very slowly until reliable assessments indicate that increased harvests are sustainable’. A condition of the EU licensing scheme is that member states must submit a sampling plan for deepwater species that includes the deployment of scientific observers on vessels. This is an important first step that will lead to increased availability of fisheries and biological data for assessments, although NEAFC have yet to decide whether or not a similar scheme should be introduced for non-EU vessels fishing in international waters. However, new fisheries will continue to be identified and developed with a minimum of regulation and control and, most importantly, without any assessment of sustainable fisheries production potential. Until these issues are resolved and real steps taken to drastically reduce fishing effort across most stocks, few deepwater fisheries in the Northeast Atlantic are likely to be biologically sustainable. If fisheries management bodies do not grasp the nettle the only major force likely to reduce the level of fishing effort will be economic, i.e. when stocks become mined out and fishing becomes economically non-viable. This pattern can already be seen in stocks of orange roughy, a species highly vulnerable to exploitation and possibly an early indicator of future trends.

6. ACKNOWLEDGEMENTS

We thank all members and participants of the ICES Study/Working Group on the Biology and Assessment of Deep-sea Fisheries Resources for their contributions in providing data and analyses used in this manuscript.

7. LITERATURE CITED

Anon. 1998. Report of the Study Group on the Biology and Assessment of Deep-sea Fisheries Resources. ICES CM 1998/ACFM:12, 172 pp.

Anon. 2000. Final report of the EU Study Project CT 97/0084. Environment and biology of deepwater species Aphanopus carbo in the NE Atlantic: Basis for it’s management (BASBLACK). DG XIV European Commission, 94 pp.

Anon. 2001. Report of the ICES Advisory Committee on Fishery Management 2000. ICES Coop. Res. Rep., 242, 911 pp.

Anon. 2002a. Report of the Working Group on the Biology and Assessment of Deep-sea Fisheries Resources. ICES CM2002/ACFM: 16,256pp. CM 2002/16, 256 pp.

Anon. 2002b. Report of the ICES Advisory Committee on Fishery Management 2001. ICES Coop. Res. Rep., 255, 948 pp.

Anon. 2003. Report of the Working Group on the Biology and Assessment of Deep-sea Fisheries Resources. ICES CM2003/ACFM: 25,41pp. CM 2003/25, 41 pp.

Bergstad, O.A., & O.R. Godø 2002. The pilot project "Patterns and processes of the ecosystems of the northern Mid-Atlantic": aims, strategy and status. Oceanologica Acta 25: 219-226.

Bergstad, O.A., & N.R. Hareide 1996. Ling blue ling and tusk of the north-east Atlantic. Fisken og Havet, 15, 1996. 126 pp.

Cochrane K.L. 2002 (Ed.). A fishery manager’s guidebook - Management measures and their application. FAO Fish. Tech. Pap. 424, 231 pp.

Connolly, P.L., & C.J.Kelly 1996. Catch and discards from experimental trawl and longline fishing in the deepwater of the Rockall Trough. J. Fish Biol., 49 (Supplement A): 132-144.

EC 2002a. Council regulation (EC) No. 2340/2002 fixing for 2003 and 2004 the fishing opportunities for deep-sea fish stocks. Official Journal of the European Communities. L 356, 11 pp.

EC 2002b. Council regulation (No 2347/2002 - Establishing specific access requirements and associated conditions applicable to fishing for deep-sea stocks. Official Journal of the European Communities. L 351, 11 pp.

Francis, R.I.C.C. 1992. Use of risk analysis to assess fishery management strategies: a case study using orange roughy (Hoplostethus atlanticus) on the Chatham Rise, New Zealand. Can. J. Fish. Aquat. Sci., 49: 922-930.

Francis, R.I.C.C. 1993. Monte Carlo evaluation of risks for biological reference points used in New Zealand fishery assessments. Can. Spec. Publ. Fish. Aquat. Sci., 120: 221-230.

Francis, R. I.C.C., M.R. Clark, R.P. Coburn, K.D. Field & P.J. Grimes 1995. Assessment of the ORH 3B orange roughy fishery for the 1994-95 fishing year. New Zealand Fisheries Assessment Research Document 95/4. 43 pp. (Unpublished report held in NIWA library, Wellington).

Garcia, S.M. 1994. ’The precautionary principle: its implications in capture fisheries management’, Ocean and Coastal Management, 22: 99-125

Gordon, J.D.M. 1998. Deepwater fish and fisheries in the northeastern Atlantic and Mediterranean: an overview of the EC FAIR Deep Fisheries Project. ICES CM 1998/O:10, 14 pp.

Gordon, J.D.M. 2001. Deepwater fisheries at the Atlantic Frontier. Cont. Shelf Res., 21: 987-1003.

Gordon, J.D.M., O.A. Bergstad, I. Figueiredo, & G. Menezes 2003. Deepwater fisheries in the Northeast Atlantic. J. Northw. Atl. Fish. Sci.

Hilborn, R., and C.J. Walters 1992. Quantitative Fisheries Stock Assessment - Choice, dynamics and Uncertainty. Chapman & Hall, London.

Koslow, J.A., G. Boehlert, J.D.M. Gordon, R.L. Haedrich, P. Lorance, & N. Parin 2000. Continental slope and deep-sea fisheries: implications for a fragile ecosystem ICES J. mar. Sci., 57(3): 548-557.

Large, P.A. 2002. The use of stock reduction to assess deepwater species in the ICES area - an initial investigation. Working paper submitted to the ICES Working Group on the Biology and Assessment of Deep-sea Fisheries Resources, 2002 (unpubl).

Large, P.A., C. Hammer, O. A. Bergstad, J. D.M. Gordon & P. Lorance 2003. Deep-water Fisheries of the Northern Atlantic: II Assessment and Management Approaches. J. Northw. Atl. Fish. Sci.

Magnússon, V., O.A. Bergstad, N.-R. Hareide, J. Magnússon & J. Reinert. 1997. Ling, blue ling and tusk of the northeast Atlantic. Tema Nord, 535: 58 pp.

Menezes, Gui, A. Rogers, H. Krug, A. Mendonça, B.M. Stockley, E. Isidro, M.R. Pinho & A. Fernandes 2001 Seasonal changes in biological and ecological traits of demersal and deepwater fish species in the Azores. Final Report, European Commission DGXIV/C/1 - Study contract 97/081.

Merrett, N.R. & R.L. Haedrich 1997 Deep-sea Demersal Fish and Fisheries. Chapman & Hall, London. 282 pp.

OSPAR 2000. Quality Status Report 2000, Region V-Wider Atlantic. OSPAR Commission, London. 110 pp.

Pechenik, L.N. & F.M. Troyanovsky 1970. Trawling resources on the North Atlantic continental slope. Israel Program for Scientific Translations, 1971, Cat. No. 5977, 66 pp.

Pope, J. 2002. Input and output controls: the practice of fishing effort and catch management in responsible fisheries. FAO Fish. Tech. Pap. 424: 75-93.

Swan, S.C., D.M. Gordon, B. Morales-Nin, T. Shimmield, T. Sawyer & A.Geffen 2003. Otolith microchemistry of Nezumia aequalis (Pisces: Macrouridae) from widely different habitats in the Atlantic and Mediterranean. J. mar. biol. Ass. UK, 83: 883-886.

Troyanovsky, F.M. & S.F. Lisovsky 1995. Russian (USSR) fisheries research in deepwaters (below 500 m) in the North Atlantic. In: Deepwater fisheries of the North Atlantic Oceanic Slope, A.G. Hopper (Ed). Kluwer Academic Publishers, The Netherlands, 357-366.

Uiblein, F., P. Lorance & D. Latrouite 2003. Behaviour and habitat utilization of seven demersal fish species on the Bay of Biscay continental slope, NE Atlantic. Mar. Ecol. Prog. Ser., 257: 223-232.

Potential exploitable deepwater resources and exploratory fishing off the South African coast and the development of the deepwater fishery on the south Madagascar ridge

D.W. Japp^[44] and A. James^[45]

1. INTRODUCTION

The evolution of deepwater fisheries in the Southern African context has been poorly documented. In this paper we present our account of the development of deepwater fishing off the Southern African continental shelf, including the high-seas demersal fishing effort in the southeast Atlantic Ocean, the eastern Indian Ocean and the area directly south of the South African coastline. We provide an opinion on the commercial availability and potential of the main deepwater species in these areas.

2. TIME FRAMES

The development of South African interests in deepwater exploitation is relatively recent. However in this report we identify convenient periods of activity in the deepwater sector off the Southern African sub-continent as follows.

• Up to 1989, including the effort prior to the introduction of the South African 200 mile fishing zone (EEZ) in 1977, the introduction of quota management of the South African hake fishery from 1978 and the disbanding of the International Commission for Southeast Atlantic Fisheries (ICSEAF) in 1989 (just prior to Namibian Independence).

• From 1989 to 1994 including the transition period in which Namibia took control of their national fisheries and the change in political dispensation in South Africa after which the process of a new fisheries policy was initiated.

• From 1995 to 2000, including the ratification of the 1982 United Nations Convention of the Law of the Sea by Namibia and South Africa with the subsequent formation of the regional fisheries agreement Southeast Atlantic Fisheries Organisation (SEAFO) and the High Seas Fisheries Agreements including the Straddling and Highly Migratory Fish Stocks agreements.

• From 2000 to present - deep-sea fishing effort in the Indian Ocean.

Each of the time frames specified above have relevance to the development of high-seas fishing in the Southern African context as they imposed different fisheries management and political regimes on national, regional and international fishing operators.

3. HIGH-SEAS EFFORT IN THE SOUTHEAST ATLANTIC UP TO 1989

Effort directed at groundfish resources both on the continental shelf and on the high seas in the Southeast Atlantic is poorly documented. South Africa extended management of their regional fisheries, through the declaration of a 200 mile EEZ in 1977. South African fishing effort was contained within its own national boundaries and little bottom-trawl effort off the continental shelf was reported. Rights holders in the bottom-trawl sector targeted mostly the Cape hakes (Merluccius capensis and M. paradoxus), the biomass of which was steadily improving under a conservative management regime, so there was no incentive to seek alternative potential resources.

Namibian fisheries on the other hand were being managed through the collective efforts of the ICSEAF signatory states of which South Africa, Spain and the USSR were prominent. The significance of this was that it permitted the introduction of distant water fleets, (particularly Spain and the Soviet Block countries) into the region. At the time there was no coordinated effort to manage the high seas adjacent to the territorial waters of Namibia (formerly South West Africa). Research in the form of swept-area biomass surveys were conducted in the ICSEAF period and South African records at this time reported occasional large incidental catches of orange roughy (Hoplostethus atlanticus) in Namibian waters. Circumstantial evidence also pointed to large catches of orange roughy been taken by Russian trawlers in the vicinity of the hake trawling grounds. In addition to large amounts of hake and other bycatch species, it is believed that these incidental catches of orange roughy were processed into fishmeal (the processing and markets for orange roughy were not yet fully developed).

In addition to the mostly unreported or misreported catches off the Namibian coast, the exploratory efforts of the Russian international fleet, known as AtlantNIRO, were obtained when evaluating the historical catch records for the proceedings leading to the formation of the Southeast Atlantic Fisheries Organisation (SEAFO) (Japp 1999). Target species were alfonsino (Beryx splendens), orange roughy (Hoplostethus atlanticus), armourhead (Pseudopentaceros richardsoni) and cardinal fishes (Epigonus spp.). The USSR exploratory fleet systematically trawled seamounts and banks along the mid-Atlantic ridge (Figure 1, reference points 1 and 2) extending southwards from the Azores along the Walvis Ridge as far as the Tristan da Chuna Island complex. Documents translated (from Russian) of these exploratory trips are explicit, describing exact locations, the nature of each feature and the identification of targets and species composition of catches. There is no doubt that large volumes of deepwater species were caught and conservative estimates are that at least 50000 tonnes was caught in the period 1975 to 1979.

4. HIGHSEAS EFFORT IN THE SOUTHEAST ATLANTIC FROM 1989 TO 1994

From a fisheries management and political perspective this was a critical period for the Southeast Atlantic. Namibia obtained independence in 1990 and immediately exerted control over their fisheries resources. Foreign fishing fleets were effectively removed from the region. Further, South Africa no longer influenced the research and management of the Namibian fisheries. At the end of the period, South Africa also moved into a new political order and initiated the development of a new fisheries policy from 1994 which took several years to establish.

Although Namibia controlled effort within their EEZ, foreign-flagged vessels operating on the high seas in waters adjacent to their border was not discouraged as utilisation by foreign vessels of Namibian ports was a much needed economic boost for the region. Reported high-seas landings were however minimal and poorly recorded (mostly small quantities of alfonsino). Towards the end of the period, increasing interest in deepwater stocks was being shown by South African operators.

Several reasons for this are suggested including

• he development of markets for deepwater species (particularly orange roughy and alfonsino)

• worldwide recognition of the potential of deepwater resources

• threats of limit on hake quotas (South Africa)

• development of deepwater fishing techniques and

• deployment of excess capacity in the hake fishery.

In June 1994 the first application was submitted to the South African fisheries management authority (Sea Fisheries Research Institute) for an exploratory deepwater fishing right. With the subsequent granting of a single exploratory deepwater permit, a deepwater vessel was deployed in the South African EEZ specifically to conduct exploratory fishing for deepwater resources. This resulted in the discovery in late 1994 of the deepwater oreo dory resource on the southern tip of the Agulhas Bank in 700-1100 m (Figure 1, reference point 4). Although considerable effort was deployed, indications were that orange roughy abundance off the South African coast was restricted to the west coast extending from the cape point northwards to the Namibian border. Catches Figure 1 (separate file) of orange roughy were however limited to very small quantities with no definite indications of aggregations or areas with the potential for larger quantities.

5. EXPLORATORY DEEPWATER FISHING OF SOUTH AFRICA AND NAMIBIA FROM 1995

5.1 The beginnings

Interest in deepwater fishing off South Africa and Namibia acquired increasing momentum in this period. Extensive research (side-scan sonar) and trawl surveys were conducted mostly by a single operator. The South African management authority issued more deepwater permits (5) to local operators between 1995-1997. In this period specific legislation was introduced prohibiting the landing of deepwater species caught inside the South African fishing zone without a permit.

5.2 Effort off South Africa

Several South African operators activated their permits. Few, however, designated boats full-time on deepwater fishing and instead only used their vessels as capacity in their hake-directed fleet became available. Some permit holders fished the Walvis Ridge, particularly Valdivia Bank. Small volumes of orange roughy and alfonsino were reported. Pelagic armourhead were also targeted and a few large landings from Valdivia were reported (Japp 1999). Within South African waters, effort continued on the oreo dory (Figure 1, reference point 4) and some effort was also directed westwards into the South Atlantic. Catches of deepwater dory were however limited to smooth dory (Pseudocyttus maculatus), warty dory (Allocyttus verrucosus) and spikey oreo (Neocyttus rhomboidalis). The fish caught were generally small by comparison, for example, to catches of similar deepwater dory species off New Zealand and Australia (Ward et al. 1996). Because of the limited area being fished, management considered of restricting effort and setting precautionary catch limits on the deepwater dory stocks.

This management never materialized. South Africa was going through an intensive period of fisheries policy development and rights-based concerns, mostly focused on transformation of the fishing industry. Under threat of losing their bread and butter (hake) resources, most trawler operators with the capital and resources capable of taking the risk of conducting costly deepwater research were reluctant to put any further effort into the fishery. Of the five deep-sea permit holders, three consolidated for a period and fished using a single vessel, while the others continued to fish intermittently.

Despite the rights-based concerns in the South African fishing industry, there were still operators interested in developing South Africa’s deepwater fisheries. The state management authority (Marine and Coastal Management) issued the last deepwater permits in 1998 and then called for applications on a more formal basis for 1999. Applications were submitted by numerous operators - these applications were however not processed and subsequent follow-up requests by potential operators have never been considered by the MCM.

The status of the South African deepwater fishery therefore remains uncertain, although the MCM have indicated that applications for a range of new and exploratory fishing rights will be called for in the near future.

5.3 Effort in Namibian waters

Advances in deepwater fishing off Namibia in this period was much more positive than in South Africa. Namibia had established a policy and management framework that encouraged exploratory fishing for new resources. The development of the Namibian orange roughy fishery is now well known and beyond the scope of this paper. In summary however, the South African operator responsible for the initial research thrust in South African waters in 1994 and 1995 identified Namibian waters as having potential and acquired an exploratory deepwater permit there. From 1995 four orange roughy grounds were identified and exploited under a management regime in Namibia (Figure 1, reference point 3). Despite management efforts and catch restrictions however, the catch rates in Namibia rapidly declined and although the fishery is still active, it is restricted to a relatively small catch with the periodic closing of several of the quota management areas (QMAs).

The development of the Namibian orange roughy fishery involved the pioneering use of side scan sonar technology. The technology was employed to identify potentially suitable orange roughy habitat - seamounts, drop offs, canyons. This technology facilitated the development of all four main fishing grounds in Namibia. Further, this technology played a major role in the development of the South West Indian Ocean Fishery.

6. DEVELOPMENT OF THE INDIAN OCEAN DEEPWATER FISHERY

From 1996 greater interest was being shown in the potential for fisheries in the high seas, particularly from established Australian and New Zealand operators some of whose vessels had conducted exploratory fishing westwards into the Southern Indian Ocean. In this regard exploratory effort was directed at areas south and east of Madagascar (Figure 1, reference point 5 and ringed areas). South African operators also began to fish more aggressively in the high seas. South Africa had introduced a new fisheries policy and the rights of many of the established companies were under threat of either removal or reduction. South Africa also encouraged high-seas development through the issuing and control of high-seas permits (a process encouraged by the United Nations Convention on the Law of the Sea and subsequent high-seas and straddling-stocks agreements). In this regard the targeting on orange roughy by South African boats on the Tasman Rise in 1999 was a wake up call not only for the South African authorities, but also in the international context with regard to bilateral arrangements between New Zealand and Australia and the management of high-seas resources adjacent to coastal states.

Prior to the South African campaign to the Tasman Rise in 1999 there had been a well organized, low profile campaign conducted by New Zealand operators in the Indian Ocean to explore for, and develop, deepwater resources in the Indian Ocean. The campaign made use of historical catch information from the Russian exploratory fleets of the 1970’s (as reported for the Southeast Atlantic) and the new side scan sonar technology that had been pioneered successfully in Namibia and New Zealand.

FIGURE 1. Development of deep-sea fishing in the Southeast Atlantic, Southern and Western Indian Oceans up to 2003

The exploration, which commenced in 1998, was extensive, covering areas from the Madagascar Ridge (including Walters Shoals), South West Indian Ocean Ridge, and the Discovery, Indomed and Galleini Fracture Zones in the west to the Ninety East Ridge and Broken Ridge in the east (Figure 1 refers). The campaign - essentially one vessel - was successful in locating commercial quantities of alfonsino and orange roughy.

Independently of the New Zealand campaign, an Australian operator was also conducting exploratory fishing in the South West Indian Ocean. As fate would have it, the New Zealand (1), Australian (2) and South African (2) vessels returning from the Tasman Rise all came upon each other in the latter half of 1999. This chance meeting on the high seas immediately sparked competition for fish off the limited grounds that had been identified in the South Indian Ocean.

The increased profile of the fishing activity in the SW Indian Ocean soon drew interest from deepwater operators around the world. The initial five vessels were soon joined by Korean vessels that had been active on the Tasman Rise. By the end of 1999 the fleet had doubled to 10-12 vessels from six or seven flag states. By mid-2000 more than 35 vessels from more than 13 flag.....flag states were reportedly operating in the South West Indian Ocean.

The fishery peaked in 2000/2001 with catches in excess of 15 000 tonnes of orange roughy being landed at various ports in Australia, Indonesia, Mauritius, Seychelles, Mozambique, South Africa and Namibia. As in the early days of the Australian and New Zealand orange roughy fisheries, there were many reports of vessels queuing to shoot their gear on productive tow lines with occasional large catches reported (Figure 2).

The labile environmental conditions on many of the seamounts being targeted resulted in catches being sporadic at times with the effect of environmental fluctuations on fish availability being exacerbated by the high level of fishing activity. This forced operators to diversify their fishing and target other species, particularly alfonsino, boarfish and bluenose.

Orange roughy catches declined significantly in 2001, and more so in 2002. The large fleet that had built up in 2000, dissipated, as rapidly as the catches declined and by 2002 the fleet comprised probably no more than 15 vessels, at the most. By mid-2003 the fleet was approximately 8. vessels, which included the original New Zealand, Australian and South African operators.

It is clear that success in this fishery comes down to experience, preparation (cf. New Zealand campaign) and the ability to ride out long periods of low catches.

7. POTENTIAL FOR THE DEVELOPMENT OF NEW DEEPWATER FISHERIES OFF SOUTH AFRICA

We conclude that the economic potential of a deepwater fishery exists off South Africa both within the economic zone and the adjacent high seas. Results to date have indicated suitable species availability, but the true economic potential needs to be investigated through renewed exploratory effort. The obvious species would be deepwater dory, although the availability of orange roughy also cannot be excluded.

Potential deepwater stocks are probably most likely to be found on the West Coast where the current, bathymetric and habitat characteristics are likely to be suited to their life history strategies. The Benguela Current system off the West Coast is, however, dominated by an upwelling regime with highly variable bottom temperatures and periodic anoxic water. These conditions are likely to affect known orange roughy (and other deepwater species) behavioural regulators (such as temperature preferences). South African operators have, however, been progressive in their approach to deepwater fishing with the development of techniques and good use of available technology (as indicated by the development of the SW Indian Ocean fishery).

The potential for deepwater resources off the South African south and east coasts is less promising with the strongly-flowing Agulhas current influencing not only the behaviour of deepwater fish species, but also making deepwater trawling difficult. Grounds, including significant amounts of deep-sea corals exist, which would complicate fishing, and the potential for deepwater species in these areas remains to be fully investigated.

We conclude therefore that fisheries management in South Africa should be more progressive and supportive of deepwater exploration. Few dedicated deepwater operators willing to commit to exploratory deepwater fishing exist in Southern Africa and in this respect, the experience and influence of those that do, as well as other supporting deepwater nations, should be recognized. Fisheries managers in South Africa have the advantage of the experience of deepwater fisheries in other parts of the world, and in particular the recent demise of the Indian Ocean resource, is one example. Deepwater fisheries are, however, costly and high-risk operations, and in this context unlikely to attract new or inexperienced South African fishers.

8. LITERATURE CITED

Japp, D.W. 1999. Updated review of the fish resources and catch statistics in the proposed SEAFO area. Unpublished report

Ward, R.D., N.G. Elliot, G.K. Yearsley and P.R. Last 1996. Species and stock delineation in Australasian oreos (Oreostomatidae). Final Report to Fisheries Research and Development Corporation (CSIRO), Grant 92/24. Division of Fisheries, 144 pp.

Counting deepwater fish: challenges for estimating the abundance of orange roughy in New Zealand fisheries

M. Clark
Deepwater Fisheries, NIWA
Private Bag 14-901, Kilbirnie, Wellington, New Zealand
<[email protected]>

1. INTRODUCTION

Orange roughy (Hoplostethus atlanticus) are an important commercial fishery species in several parts of the world. These fisheries are comparatively recent, developing around New Zealand in the late 1970s, and subsequently in Australia, the North Atlantic off the Faroes and Scotland, Namibia, Chile, and the Indian Ocean (Francis and Clark 2004).

Orange roughy is a deepwater species, occurring at depths between 500m and 1 500 m. Because of this, they cannot be kept alive easily and studied in tanks, or tagged and returned. The water conditions in which they live feature high pressure and low temperature so aspects of their biology and ecology are still not well understood. They can form very dense, and predictable, aggregations, and these can be over a range of habitat - from flat bottom to canyon edges to the tops and sides of hills, and importantly, often over rough ground. The species can be highly vulnerable to overfishing, but slow to recover from it. It means that estimating abundance is critical to successful management, especially in the early years when the virgin stock can be rapidly fished down.

General methods of determining fish abundance have been described frequently and reviewed in fisheries texts and journals (e.g. Saville 1977, Sissenwine, Azarovitz and Suomala 1983), but there have been few published accounts of survey results for deepwater species like orange roughy. Clark (1996a) described some elements of performance of abundance estimation methods applied to deepwater species in New Zealand, and Branch (2001) extended this with Namibian fishery results. Since then, data and technology improvements have occurred, and the conduct, or understanding of the efficacy, of some methods have developed.

In this paper, the work of Clark (1996a) is updated. Four methods used in measuring stock size of orange roughy in New Zealand waters are briefly described and their effectiveness discussed: trawl surveys, acoustic surveys, egg production surveys and analysis of commercial catch and effort data. Combinations of methods, and indirect estimation procedures are also considered. The intention in this paper is not to describe these methods in detail, as the general methodology is the same as that used for shallow water species and is covered by Clark (1996a). Rather, the aim is to focus on specific features and aspects of the methods relevant to the particular characteristics of orange roughy and similar deepwater species.

2. TRAWL SURVEYS

2.1 Survey design

Swept-area trawl survey methodology is widely used and well documented (e.g. Grosslein 1979, Doubleday and Rivard 1981). In most New Zealand deepwater surveys, a two-phase stratified random design has been employed (after Francis 1984). Trawls are carried out random positions, or a combination of position and depth, parallel to the depth contour, over a fixed distance. These tows are repeated each year in some surveys, or new random stations are generated for each survey. In areas where fish occur on seamount or ‘drop-off’ features, modifications are made to the standard survey design (Clark 1994, 1996a). Stratification is typically based on narrow depth bands, and is also done around such seamount features, because orange roughy form dense aggregations in localized areas.

In the early years of fishery development, a guide to the level at which to set catch limits was urgently needed by managers. Swept-area estimates have been used as absolute biomass, despite poor knowledge of fish herding by, or escapement from, the trawl gear. As time series were built up, indices were used as measures of relative abundance (Francis 1992, Clark 1995).

2.2 Trawl survey descriptions

Time series of trawl surveys have been carried out in five New Zealand orange roughy fisheries: the Chatham Rise ‘spawning box’, Challenger Plateau, East coast of New Zealand, Puysegur Bank and Bay of Plenty. Locations of the main fishing grounds are shown in Figure 1.

The Chatham Rise ‘spawning box’ has been the most exploited and researched orange roughy fishing ground. Trawl surveys were the main method of monitoring abundance over the period from 1984 to 1994 (Anderson and Fenaughty 1996). Overall, the indices from the Chatham Rise surveys through the 1980s showed a consistent declining trend. The variance of the survey estimates was generally low and the trend in indices was consistent with an observed contraction in the distribution of fish concentrations (Clark et al. 2000), and together with a relatively good fit of the indices to orange roughy computer models, gave confidence in the estimates of biomass.

However, surveys in 1992 and 1994 showed increasing levels of uncertainty in the biomass index, as the coefficient of variation (CV) increased from typically less than 20 percent in the 1980s to 34 percent in 1992 and 67 percent in 1994. The situation of a shrinking area of high fish density is consistent with an increasing CV, as in using the same stratum boundaries there is an increase in the ‘hit-or-miss’ situation. The area was also closed to commercial fishing in 1992 and this could have lead to an increase in fish concentration as heavy fishing pressure may disrupt the formation or stability of aggregations (Clark and Tracey 1991). The differing results in 1992 and 1994 from previous surveys appear related to a changing pattern of fish distribution, relative abundance and availability. The trawl survey technique is thought to have worked well at tracking abundance during the 1980s, but started to fail with the changing fish distribution in the early 1990s. The need for tight stratification when small concentrations occur, or a large number of tows (many with high catch rates) in a short period when the distribution of fish is stable, limits the application of the technique in the present situation. Thus, the time series has now been discontinued.

Surveys in other areas have also shown strong trends. The Challenger Plateau winter survey indices decreased dramatically between 1987 and 1989, which was a time when commercial catch rates also dropped and the fishery showed signs of overexploitation (Clark and Tracey 1994). Similarly, in the Puysegur Bank region there have been substantial decreases in indices between consecutive surveys (Clark 1996a), although different vessels and timing limit the usefulness of the comparisons. Results from non-winter trawl surveys off the east coast have been difficult to interpret. Surveys between 1992 and 1994 showed little contrast in the index values, despite commercial catch rates in the main fishing season decreasing substantially and a subsequent stock assessment showing the period to be one where the stock size had fallen (Field et al. 1994). Variance was still reasonably high, and it is possible that the wide area survey was mainly tracking consistently low densities of dispersed fish, with an occasional catch from pockets of feeding aggregations scattered on hill and drop-off features through the area. Trawl surveys in the western Bay of Plenty covered several hill features where orange roughy were known to spawn, and where most of the catch was taken in the 1990s (Clark 1996b, Clark and Field 1998). The time series was established as part of an adaptive management programme (Starr, Clark and Francis 1996). Survey indices showed a strong decline, and although there is uncertainty about a possible change in environmental conditions affecting the fish, there is little sign of any improvement, and the area is currently closed to targeted orange roughy fishing

Trawl surveys have advantages over a number of other fishery-independent methods. They do not require a specific research vessel, or highly specialized scientific equipment, there is a well-established methodology and survey design, and results can be produced quickly. However, conditions of bottom type and fish distribution must be suitable for the technique to work. Orange roughy often aggregate over rough seafloor, and at high densities such that rapid gear saturation can occur and not provide an accurate catch rate measure.

3. ACOUSTIC SURVEYS

3.1 Survey methods

Acoustic techniques have been widely used in fisheries research. Because orange roughy can form dense schools, with little mixing with other species, they are potentially well suited to estimation of abundance by acoustic methods.

General acoustic methods are described in detail by Johanesson and Mitson (1983) and MacLennan and Simmonds (1992). Survey design in New Zealand on relatively flat sea bottoms has been based on the vessel following parallel transects within strata, usually with random spacing of the transects (Jolly and Hampton 1990) and a "starburst" pattern over hill features (Doonan, Bull and Coombs 2003a).

3.2 Survey descriptions

Major acoustic surveys of orange roughy have been carried out by NIWA in three areas:

These surveys were all of spawning aggregations in the winter months of June or July. They had been treated as absolute estimates until time series were developed. Acoustic techniques are a preferred method in New Zealand because of the dense aggregations that characterize most fishing grounds. However, there are still a number of significant issues to resolve.

Orange roughy have a swimbladder filled with wax esters rather than gas, and this gives a relatively low target strength (Do and Coombs 1989, Elliot and Kloser 1993). There has been considerable debate, and work, directed at the target strength of orange roughy (e.g. Kloser, Williams and Koslow 1997, McClatchie et al. 1999), but there is now some concordance amongst researchers that an average tilt-adjusted target strength value for a 35cm SL fish is about -50 dB (McClatchie and Coombs 2000, Kloser et al. 2000). Changes in the assumed value of target strength (for both orange roughy and other species in lower density areas) can make appreciable differences to the resulting biomass estimates. Doonan et al. (2003) calculated that a 3dB change in target strength values for orange roughy could alter the abundance estimate for a survey off the east coast of the North Island in 2001 by about 20 percent, and the same change for bycatch species caused a shift of 40-60 percent.

A further difficulty encountered with the acoustic method used for New Zealand orange roughy is the frequent distribution of the fish in areas of steep slope, canyon edges, or the sides of seamount features. Extensive ‘bottom-shadowing’ can occur, resulting in an acoustic dead-zone close to the bottom. When the slope of the bottom is about 15 ° (common on hill features), the transducer needs to be within about 200 m of the bottom to reduce the dead-zone height to less than 10s of metres. A deep-towed body (capable of being towed at 700-800 m) and digital acoustic system, is needed to do this.

The identification of the species composition from echosounder marks is a task that can require an appreciable amount of support trawling. Aggregations of orange roughy typically have a small proportion of bycatch species. However, the vertical extent needs careful definition, and there may be appreciable mixing off the bottom with mesopelagic fish and zooplankton. Outside the aggregations, orange roughy are widely distributed, and a substantial biomass can occur though at low densities. Estimates from NIWA surveys of areas of the Chatham Rise have indicated that as much as 60 percent of the stock may occur outside the main aggregations during the winter spawning season (Bull et al. 2000, I. Doonan, NIWA, Wellington, New Zealand, pers. comm.). In these cases, separation of orange roughy from other species is important for reliable measurement of biomass. The target strength of these other species is often poorly known, but they are generally higher than orange roughy. Therefore, a small number of other species can dominate the backscatter and swamp that of orange roughy. Use of multiple frequencies (Kloser et al. 2000), phase-change and chirping techniques (Barr and Coombs 2001) are recent advances in species discrimination that can overcome some of the inadequacies of using the species mix from trawl catches for interpreting acoustic data (Kloser, Williams and Koslow 1997).

Weather can also limit the use of acoustic assessment methods. Rough sea conditions can cause an extensive bubble layer to form under the vessel (and its acoustic transducer) as it pitches and rolls. This attenuates acoustic signal and results in a broken data record that makes it difficult to identify the seafloor and distinguish fish echos. In relatively calm seas, over flat bottom, hull-mounted systems may be satisfactory, and they may also be acceptable in worse weather if running with a following wind or swell. However, a towed body may be needed to get the transducer below the sea surface and clear of such disturbance. These towed bodies containing the transducer can still be towed relatively quickly, at depths of 50-100 m below the surface.

4. EGG PRODUCTION METHODS

4.1 Survey methods

There are two egg production methods which have been used to estimate spawning biomass of orange roughy.

i. Annual egg production (AEPM) (Saville 1964, Picquelle and Megrey 1993) This method estimates biomass by dividing the annual egg production in the survey area (which is the sum of daily planktonic egg production estimates made through the spawning season) by the product of the weight-specific annual fecundity of the females and the proportion of females.

ii. Daily fecundity reduction (DFRM) (Lo et al. 1992, Lo et al. 1993) Biomass of spawning females is derived from division of the daily planktonic egg production in the survey area by the weight-specific daily fecundity of females.

4.2 Survey description

Egg production surveys have been carried out on three New Zealand orange roughy spawning grounds: Ritchie Banks (in 1993 and 1995) East Cape (in 1995) and Northwest Chatham Rise (in 1996). In all areas, the DFRM was the main method used. Details of the surveys are given by Zeldis et al. (1997) and Francis, Clark and Grimes (1997) and are summarized by Clark (1996a).

Several important problems were identified during New Zealand egg surveys and subsequent analysis of data. Advection of older eggs out of the survey area often limited the range of egg ages able to be included in the analysis. Original survey boundaries were sometimes not large enough. Extensive hydrographic work needs to be carried out before the egg survey to measure current direction and speed, and enable estimates of egg drift to be made, aiding definition of the extent of the survey area and appropriate strata.

On one of the fishing grounds there was possible turnover of fish during the survey, which affects the applicability of the DFRM method. There was evidence that fish were leaving the spawning area before spawning had finished, which meant that spent fish were under-represented in trawl samples by about 40 percent.

A major problem was encountered with analysis of data from East Cape and Northwest Chatham Rise surveys in that the estimate of egg mortality (Z) from several subsurveys completed were negative (i.e. younger eggs were less abundant than older eggs) (Francis, Clark and Grimes 1997). The extremely localized production of eggs typical of orange roughy spawning grounds is generally of benefit for egg production surveys, but in this case it appears that the release of eggs was from such a small area, that within the central stratum younger eggs were sampled less frequently than the older eggs which had become more dispersed. Temporal patchiness in spawning (i.e. pulses of egg release) also seems to have contributed to the problem. To resolve this frequent and intensive sampling was required in the central stratum.

Egg production techniques have the potential of giving a measure of absolute biomass. However, because the distribution of orange roughy eggs is patchy, results typically have a high variance. In addition, the survey design and data analysis can be complex. The AEPM approach is likely to be more robust than DFRM, in that turnover does not affect results. However, AEPM surveys require sampling to occur over most of the spawning period, which can involve 4-6 weeks of vessel time. Specialized equipment and experienced staff are needed. But, there is a fine balance between the need to sample the core area intensively (to minimize the problem of negative estimates of total mortality Z) yet cover a sufficiently large area to sample older eggs before they are carried by advection beyond survey boundaries. Trawling during the survey, and subsequent examination and analysis of fecundity data, is also relatively time consuming and resource intensive. An additional trawl survey is needed prior to the egg survey to estimate the sex ratio, and proportion of females that are not going to spawn that year and hence not join the spawning aggregations. These factors make the technique relatively expensive.

5. COMMERCIAL CATCH AND EFFORT DATA

5.1 Data and analysis

Catch and effort data are routinely collected from commercial fishing operations in New Zealand. In standardized analysis of orange roughy CPUE, multiple regression techniques are used, Variables such as fishing year, season, area, nationality, vessel size, and vessel power are regressed against the log of CPUE (generally catch per tow or catch per vessel day). The model incorporates all the factors affecting catch rate and the resultant year effect is used as an annual abundance index.

A generalized linear model (GLM) is fitted following Allen and Punsley (1984) as modified by Doonan (1991), Vignaux (1992, 1994), and Francis (2001a). Variables that best predict catch rate are selected by a forward stepwise procedure. At each step, the predictor which produces the maximum decrease in the Akaike Information Criterion (Akaike 1974) is added to the model.

A feature of some orange roughy fisheries is their "hit-or-miss" aim-trawling nature, and up to 30 percent of targeted tows may record no orange roughy catch. In such cases two separate GLMs can be fitted: a "normal" model for tows with non-zero catches, and a "binomial" model for all tows (estimating the probability of a non-zero catch). These two models are then combined multiplicatively.

The robustness of CPUE analyses is often affected by empty or missing data quota cells. As more variables that could affect catch rate are included, the more incomplete the matrix becomes. Comparability between years can be affected as effort levels decrease with quota reductions, and the fishing effort is more concentrated at certain times of the year. Often, seasonal effects are poorly estimated, and so analyses are carried out on subsets of the data, such as that for winter-only (e.g. Clark and Anderson 2003, O’Driscoll 2003, Anderson 2003).

5.2 Description of results

Regression analysis of CPUE data generally explains between 20 and 50 percent of the variance. The variance of the CPUE indices is poorly known, but thought to be relatively high. In New Zealand stock assessments, such data are typically assigned a CV of 30 percent.

Strong trends have been evident in CPUE of most New Zealand orange roughy fisheries. A commonly-observed pattern is that CPUE indices are relatively high in the first few years of a fishery, and then decrease sharply to low levels. Relative to the modelled stock trajectory, indices are above the line early, and below the line in later years. This "hyperdepletion" is demonstrated in Figure 2 for the Challenger Plateau. It is currently being investigated in greater detail through a CPUE meta-analysis (A. Hicks, University of Washington, Seattle, USA, pers. comm.). This is likely to be caused in part by the ability of fishers to target aggregations successfully and maintain high catch rates, especially during spawning in winter, even though stock size is decreasing.

Monitoring of commercial catch and effort data is an important aspect of assessing changes in most orange roughy fisheries. However, it is felt that CPUE may not track abundance accurately, although it does give valuable information on general trends in stock size.

A further important feature of orange roughy distribution relevant to CPUE analysis is that fish in some areas aggregate strongly on small seamount-type features. With increasing coverage and availability of Global Positioning System (GPS) navigation in the late 1980s, the ability of fishers to locate and direct fishing operations on such features increased. Thereby, the efficiency and effective fishing power of vessels increased through technology. Improvements in trawl gear are also a common problem in CPUE analysis and in New Zealand the type of net and ground-gear used has changed between the 1980s and late 1990s.

Changes in vessel composition of the fleet can further complicate the use of CPUE data. Individual vessel factors are an important part of standardized analyses. However, different skippers on the same vessel are widely thought to have a major bearing on fishing success. Skippers move between vessels regularly, and so it is a difficult aspect to incorporate. In recent years, quotas have been reduced, and this has changed the number and type of vessels in the fisheries, as well as their fishing patterns. The mode of fish processing has generally changed in the larger vessels from head-and-gut to fillet production. This affects the "target" catch rate, as the factory operation producing fillets at sea is slower than a head-and-gut operation, and is served by smaller catches. These all add to further confuse direct comparability of CPUE data between years.

6. COMBINED METHODS

There is no single method that stands out as being the magic answer for measuring abundance given the wide range of situations in which orange roughy fisheries occur around New Zealand. All have advantages and disadvantages, depending on the characteristics of the individual fishing grounds and fisheries (Table 1). For example, the aggregation behaviour of roughy is a problem for trawl and CPUE techniques, but facilitates acoustic and egg production surveys. Great depths can make acoustics surveys expensive where deep-towed systems are needed, and also limits tagging as an assessment method. Rough bottom can preclude trawling (both research trawl survey and commercial fishing), steep slopes can be difficult for acoustic methods, and so on. Table1 does not try to be a comprehensive pros-and-cons summary, but serves to demonstrate that no single method is a complete solution.

Table 1
Summary of some of the factors that can affect the suitability or efficiency of various methods to estimate abundance of deepwater species

(+ = positive factor for use of that method, - = negative factor)

Where resources and data permit, improved (or at least more confident) estimates of biomass can be obtained by combining results of several techniques, and most New Zealand stocks have several data sources (Table 2). The indices from these different methods are treated as separate inputs, and are not combined into any single multiple-source index, because we have no understanding of how the various measures can reasonably and appropriately be combined.

Table 2
Data sources used in combination for stock assessment of New Zealand orange roughy

The stock assessment for the Northeast Chatham Rise fishery (Francis 2001b) used abundance indices from CPUE, trawl surveys, and acoustics estimates (Figure3). The indices are in broad agreement and removing one estimate makes little difference to the estimated status of the stock. Concerns raised about individual methods were reduced by the consistency of the multiple data sources. However, in contrast, a recent stock assessment of the Mid-East Coast fishery (Anderson, Francis and Hicks 2002) included CPUE, egg survey, acoustic survey, and trawl survey indices. Estimates of biomass were relatively similar for a number of model runs where one data source was excluded, but there was an appreciable difference between results that excluded the acoustic estimate (low biomass) or excluded the CPUE (higher biomass) factor (Table 3). This raised the issue of the relative weightings given to a long relative time series (CPUE) versus a single absolute estimate (acoustics-based). No consensus has been reached on this, although a conservative management approach has been taken until further survey work has been done.

7. INDIRECT METHODS

Where there is no information on stock size, from either research survey or commercial catch-effort data, then experience from other fisheries may be useful. A "Seamounts Meta-analysis" has been carried out in New Zealand (Clark, Bull and Tracy 2001) where physical attributes and catch data of deepwater fisheries were compiled for 77 seamounts in the New Zealand region. Characteristics of location, depth, size, elevation above the seafloor, age, continental association, geological origin, distance offshore and from surrounding seamounts, and degree of spawning were defined. These were then analysed as independent variables against the minimum orange roughy population size estimated from the historical level of catch taken from seamounts to investigate whether they could be useful predictors of likely safe catch from newly found seamounts. Seamount location (latitude), geological association (whether continental or oceanic), depth of peak and slope were all useful explanatory variables in predicting likely orange roughy biomass. However, a secondary result of the study was that the biomass of orange roughy associated with any single seamount feature was low, between zero and 15000 t, with most less than 3000 t. This means that the long-term sustainable yield from such features is of the order of 300 t, or less, a year.

Table 3
Percentage change in abundance estimates of orange roughy with changes in assumed values of target strength of orange roughy and bycatch species (from Table 11 of Doonan et al. 2003)

8. DISCUSSION

Measurement of fish biomass is a difficult task. Deepwater species provide additional challenges to the normal techniques used in measuring abundance, due to their depth distribution and aspects such as aggregating behaviour and distribution in areas of steep bathymetry.

Four major methods of measuring biomass have been used in New Zealand orange roughy fisheries. Others have been tried (e.g. stereo camera techniques for calculating fish density) but have been unsuccessful. Each of the methods described in this paper has advantages and disadvantages, depending on the characteristics of the particular fish stock under study and the resources available to carry out research.

Trawl survey and CPUE methods have been applied to produce relative estimates and time series of data have been used in stock reduction analyses to estimate true biomass. Egg production and acoustic techniques have given absolute estimates. Attempts to convert trawl area-swept catch rates into absolute abundance have been unsuccessful in New Zealand (e.g. Clark 1995, Francis and Clark 2004) and Namibia (Branch 2001), and absolute estimates from acoustics are also uncertain (e.g. Boyer and Hampton 2001, Boyer et al. 2001, Bull et al. 2000, Kloser, Williams and Koslow 1997).

In general, it appears abundance is best measured by time series, which will generate relative abundance indices. This, however, assumes the gear used performs in the same way each year, which may not always be true. Characteristics of distribution and abundance of orange roughy can change over time. A technique might track abundance well for several years, but become limited if aggregation patterns change. This appears to have occurred on the Chatham Rise. We have little understanding of outside influences (e.g. environmental fluctuations, recruitment levels or variability) that could cause availability or catchability to vary between years, and this can affect our interpretation of changes in biomass indices between years. However, use of data in a relative sense is probably at this stage of our knowledge more justifiable than the large uncertainties that would exist from assuming we know how to correct our trawl performance or acoustic or egg survey measurements into true biomass. Wherever possible, a combination of methods is desirable.

With new and developing orange roughy fisheries, careful examination of all available information on habitat type, fish distribution patterns and characteristics is needed to ensure an appropriate technique is applied to estimate abundance. Ideally, an absolute measure would be available at the start of a fishery, so that catch levels in the fishing-down phase, and the longer-term target catch, can be planned. However, with deepwater species, it is particularly difficult to measure biomass from a single survey, and a method using fishery-independant relative indices might be required. The high vulnerability of species like orange roughy to overfishing, and their subsequent slow recovery, mean that development of new fisheries should be carefully controlled and surveys to measure abundance should be undertaken as early as possible. Typically, orange roughy stocks have proven smaller than originally hoped-for, or believed, and the rate of development of, and capital investment in, the fisheries has often been too high. Early management must be conservative and precautionary.

9. ACKNOWLEDGEMENTS

I acknowledge the large team of people at NIWA who over the years have worked on the biology, ecology, and stock assessment of orange roughy, and in particular Chris Francis. Most surveys, and assessments, of orange roughy on which the results and opinions in this paper are based, have been funded by the Ministry of Fisheries.

10. LITERATURE CITED

Akaike, A. 1974. A new look at the statistical model identification. IEEE Transactions on Automatic Control AC-19: 716-723.

Allen, R., & R. Punsley 1984. Catch rates as indices of abundance of yellowfin tuna, Thunnus albacares, in the eastern Pacific Ocean. Inter-American Tropical Tuna Commission bulletin 18, 303-379.

Anderson, O. 2003. CPUE analysis and stock assessment of the East Cape hills (ORH 2A North) orange roughy fishery for 2003. New Zealand Fisheries Assessment Report 2003/24. 20 pp.

Anderson, O.F. & J.M. Fenaughty 1996. Trawl surveys of orange roughy on the Chatham Rise, 1984-92. New Zealand Fisheries Data Report No. 81.

Anderson, O.F., R.I.C.C. Francis & A. Hicks 2002. CPUE analysis and assessment of the Mid-East Coast orange roughy stock (ORH 2A South, 2B, 3A) in 2002. New Zealand Fisheries Assessment Report 2002/56. 23 pp.

Barr, R. & R.F. Coombs 2001. In situ target strength measurements and chirp responses of orange roughy (Hoplostethus atlanticus). Final Research Report for Ministry of Fisheries Research Project ORH1999/01A. 45pp. (Unpublished report available from Ministry of Fisheries, Wellington)

Branch, T.A. 2001. A review of orange roughy Hoplostethus atlanticus fisheries, estimation methods, biology and stock structure. In: Payne, A.I.L., Pillar, S.C., & Crawford, R.J.M. (Eds). A decade of Namibian Fisheries Science. South African Journal of Marine Science 23: 181-203.

Boyer, D.C. & I. Hampton 2001. Development of acoustic methods for assessment of orange roughy Hoplostethus atlanticus biomass off Namibia, with specific reference to biases inherent in deepwater acoustic surveys. In: Payne, A.I.L., Pillar, S.C., & Crawford, R.J.M. (Eds). A decade of Namibian Fisheries Science. South African Journal of Marine Science 23: 223-240.

Boyer, D.C., C.H. Kirchner, M.K. McAllister, A. Staby & B. Staalesen 2001. The orange roughy fishery of Namibia: lessons to be learned about managing a developing fishery. In: Payne, A.I.L., S.C. Pillar & R.J.M. Crawford (Eds). A decade of Namibian Fisheries Science. South African Journal of Marine Science 23: 205-221.

Bull, B., I. Doonan, D. Tracey & R. Coombs 2000. An acoustic estimate of orange roughy abundance on the Northwest Hills, Chatham Rise, June-July 1999. New Zealand Fisheries Assessment Report 2000/20. 36 pp.

Clark, M.R. 1994. Aspects of random trawl survey design for deepwater species on pinnacle and drop-off features. MAF Fisheries Greta Point Internal Report 227 (unpublished report, held in NIWA library, Wellington).

Clark, M. 1995. Experience with management of orange roughy (Hoplostethus atlanticus) in New Zealand waters, and the effects of commercial fishing on stocks over the period 1980-1993. In Deepwater fisheries of the North Atlantic oceanic slope (Hopper, A.G., Ed), 251-266. Netherlands, Kluwer Academic Publishers.

Clark, M.R. 1996a. Biomass estimation of orange roughy: a summary and evaluation of techniques for measuring stock size of a deepwater fish species in New Zealand. Journal of Fish Biology 49 (suppl. A): 114-131.

Clark, M.R. 1996. Bay of Plenty orange roughy. Seafood New Zealand 4(2): 42-43.

Clark, M.R., & O.F. Anderson 2003. The Louisville Ridge orange roughy fishery: an analysis of commercial catch-effort data and stock assessment of the fishery to the end of the 2000/01 fishing year. New Zealand Fisheries Assessment Report 2003/3. 26 pp.

Clark, M.R., & K.D. Field 1998. Distribution, abundance, and biology of orange roughy in the western Bay of Plenty: results of a trawl survey, June 1995 (SMT9501). NIWA Technical Report 14.

Clark, M.R. & D.M. Tracey 1991. Trawl survey of orange roughy on the Challenger Plateau, July 1990. New Zealand Fisheries Technical Report 26.

Clark, M.R. & D.M. Tracey 1994. Changes in a population of orange roughy, Hoplostethus atlanticus, with commercial exploitation on the Challenger Plateau, New Zealand. U.S. Fishery Bulletin 92, 236-253.

Clark, M.R., B. Bull & D.M. Tracey 2001. The estimation of catch levels for new orange roughy fisheries on seamounts: a meta-analysis of seamount data. New Zealand Fisheries Assessment Report 2001/75. 40 pp.

Clark, M.R., O.F. Anderson, R.I.C.C. Francis & D.M. Tracey 2000. The effects of commercial exploitation on orange roughy (Hoplostethus atlanticus) from the continental slope of the Chatham Rise, New Zealand, from 1979 to 1997. Fisheries Research 45(3):217-238.

Do, M.A., & R.F. Coombs 1989. Acoustic measurements of the population of orange roughy (Hoplostethus atlanticus) on the north Chatham Rise, New Zealand, in winter 1986. New Zealand Journal of Marine and Freshwater Research 23, 225-237.

Doonan, I. 1991. Orange roughy fishery assessment, CPUE analysis - linear regression, NE Chatham Rise 1991. New Zealand fisheries assessment research document 91/9 (unpublished report, held in NIWA library, Wellington).

Doonan, I.J., B. Bull & R.F. Coombs 2003a. Star acoustic surveys of localised fish aggregations. Ices Journal of Marine Research 60: 132-146.

Doonan, I.J., A.C. Hicks, R.F. Coombs, A.C. Hart & D.M. Tracey 2003. Acoustic estimates of the abundance of orange roughy in the Mid-East coast fishery, June-July 2001. New Zealand Fisheries Assessment Report 2003/4. 22 pp.

Doubleday, W.G. & D. Rivard (Eds) 1981. Bottom trawl surveys: proceedings of a workshop held at Ottawa, November 12-14, 1980. Canadian Special Publication of Fisheries and Aquatic Sciences 58.

Elliot, N.G. & R.J. Kloser 1993. Use of acoustics to assess a small aggregation of orange roughy, Hoplostethus atlanticus (Collett) off the eastern coast of Tasmania. Australian Journal of Marine and Freshwater Research 44, 473-482.

Field, K.D. & M.R. Clark 1996. Assessment of the ORH 7A orange roughy fishery for the 1996-97 fishing year. New Zealand Fisheries Assessment Research Document 96/20. (Draft report held in NIWA Greta Point library, Wellington).

Field, K.D., R.I.C.C. Francis, J.R. Zeldis & J.H. Annala 1994. Assessment of the Cape Runaway to Banks Peninsula (ORH2A, 2B, and 3A) orange roughy fishery for the 1994- 95 fishing year. New Zealand fisheries assessment research document 94/20 (unpublished report, held in NIWA library, Wellington).

Francis, R.I.C.C. 1984. An adaptive strategy for stratified random trawl surveys. New Zealand Journal of Marine and Freshwater Research 18, 59-71.

Francis, R.I.C.C. 1992. Use of risk analysis to assess fishery management strategies: a case study using orange roughy (Hoplostethus atlanticus) on the Chatham Rise, New Zealand. Canadian Journal of Fisheries and Aquatic Sciences 49, 922-930.

Francis, R.I.C.C. 2001a. Orange roughy CPUE on the South and East Chatham Rise. New Zealand Fisheries Assessment Report 2001/26. 30 pp.

Francis, R.I.C.C. 2001b. Stock assessment for 2001 of orange roughy on the northeast Chatham Rise. New Zealand Fisheries Assessment Report 2001/41. 32 pp.

Francis, R.I.C.C. & M.R. Clark 2004. Sustainability issues form orange roughy fisheries. Bulletin of Marine Science.

Francis, R.I.C.C., M.R. Clark & P.J. Grimes 1997. Calculation of the recruited biomass of orange roughy on the northwest Chatham Rise using the 1996 Graveyard Hill egg survey (TAN9608). NZ Fisheries Assessment Research Document 97/29. 18 p. (unpublished(unpublished report, held in NIWA library, Wellington).

Grosslein, M.D. 1979. Groundfish survey program of BCF, Woods Hole. Commercial Fisheries Review 31(8-9), 22-35.

Jolly, G.M. & I. Hampton 1990. A stratified random design for acoustic surveys of fish stocks. Canadian Journal of Fisheries and Aquatic Sciences 47: 1282-1291.

Johannesson, K.A. & R.B. Mitson 1983. Fisheries acoustics - a practical manual for aquatic biomass estimation. FAO Fisheries Technical Paper 240.

Kloser, R.J., A. Williams & J.A. Koslow 1997. Problems with acoustic target strength measurements of a deepwater fish, orange roughy (Hoplostethus atlanticus Collett). ICES Journal of Marine Science 54: 60-71.

Kloser, R.J., T.E. Ryan, A. Williams & M. Soule 2000. Development and implementation of an acoustic survey of orange roughy in the Chatham Rise spawning box from a commercial factory trawler, F.V. Amaltal Explorer. (unpublished report available CSIRO, Hobart).

Lo, N.C.H., J.R. Hunter, H.G. Moser & P.E. Smith 1992. The daily fecundity reduction method: a new procedure for estimating adult fish biomass. ICES Journal of Marine Science 49, 209-215.

Lo, N.C.H., J.R. Hunter, H.G. Moser, & P.E. Smith 1993. A daily fecundity reduction method of biomass estimation with application to Dover Sole Microstomus pacificus. Bulletin of Marine Science 53, 842-863.

MacLennan, D.N. & J.E. Simmonds 1992. Fisheries acoustics. Chapman &

McClatchie, S. & R. Coombs 2000. In situ target strength measurements of orange roughy (Hoplostethus atlanticus). Final Research Report for Ministry of Fisheries Research Project ORH1998/01. (Unpublished report available from Ministry of Fisheries, Wellington.)

McClatchie, S., G. Macaulay, Z. Ye, R.F. Coombs, P. Grimes & A. Hart 1999. Target strength of an oily, deepwater fish, orange roughy (Hoplostethus atlanticus). 1. Experiments. Journal of the Acoustic Society of America 106: 131-142.

O’Driscoll, R.L. 2003. Catch-per-unit-effort analysis of orange roughy fisheries outside the New Zealand EEZ: Lord Howe Rise and Northwest Challenger Plateau to the end of the 2001/02 fishing year. New Zealand Fisheries Assessment Report 2003/36. 38 pp.

Picquelle, S.J. & B.A. Megrey 1993. A preliminary spawning biomass estimate of walleye pollack, Theragra chalcogramma, in the Shelikof Strait, Alaska, based on the annual egg production method. Bulletin of Marine Science 53, 728-749.

Saville, A. 1964. Estimation of the abundance of a fish stock from egg and larval surveys. Rapports et Procès-verbaux des Réunions du Conseil International pour l’Exploration de la Mer 155, 164-173.

Saville, A. (Ed) 1977. Survey methods of appraising fishery resources. FAO Fisheries Technical Paper 171.

Sissenwine, M.P., T.R. Azarovitz & J.B. Suomala 1983. Determining the abundance of fish. In: Experimental biology at sea (MacDonald, A.G.; Priede, I.G., Eds). pp. 51-101.

Starr, P., M. Clark & C. Francis 1996. ORH1 - blueprint for controlled development of an orange roughy fishery? Seafood New Zealand 4(2): 29-31.

Vignaux, M. 1992. Catch-per-unit effort (CPUE) analysis of the hoki fishery. New Zealand Fisheries Assessment Research Document 92/14. (Draft report held in NIWA Greta Point library, Wellington.)

Vignaux, M. 1994. Catch-per-unit effort (CPUE) analysis of west coast South Island and Cook Strait spawning hoki fisheries, 1987-93. New Zealand Fisheries Assessment Research Document 94/11. (Draft report held in NIWA Greta Point library, Wellington)

Zeldis, J.R., R.I.C.C. Francis, M.R. Clark, J.K.V. Ingerson, P.J. Grimes & M. Vignaux 1997. An estimate of orange roughy, Hoplostethus atlanticus, biomass using the daily fecundity reduction method. Fishery Bulletin 95: 576-597.