Catch monitoring by fisheries observers in the United States and Canada
William A. Karp1 and Howard McElderry2
1 Alaska Fisheries Science Center, National Marine Fisheries Service, 7600 Sand Point Way Northeast, Seattle, WA 98115, USA.
Email: [email protected]
2 Archipelago Marine Research Limited, #200 - 525 Head Street, Victoria, BC V9A 5S1, Canada.
Email: [email protected]
Abstract: Catches may be monitored by observers to provide data for estimating removals of target and non-target species, for stock assessment and biological research, and for evaluating compliance with fisheries regulations. While a number of methods are available for monitoring vessel activity and accounting for catch, independent, verifiable information on catch, bycatch, and discard can be provided only by at-sea observers. Thus, observer programmes are essential components of integrated fishery monitoring systems. Programmes can be successful only if their objectives are realistic, the resources they are provided with are adequate, their organisational structure allows for collection of independent, verifiable information, and the observers themselves receive the necessary training and support. Furthermore, fishing industry personnel, observer providers, and agency staff must work together to ensure effective programme operations. In this review, we examine various aspects of catch monitoring by observers with reference to issues identified during a Canada/U.S. observer programme workshop, held in 1998, and lessons learned from the operation of two large observer programme: the North Pacific Groundfish Observer Programme, which monitors groundfish fisheries off Alaska, and the British Columbia Trawl Fishery Observer Programme. We identify the factors which must be taken into account during the design, implementation, and operation of observer programmes and offer recommendations for policy makers and managers involved in operating existing programmes or developing new ones.
Commercial fishery catches are monitored for a variety of purposes. While specific objectives of monitoring programmes will depend upon the type and magnitude of the fishery, its temporal and spatial scope, and the strategies employed by its managers, there is always a need for accurate information on catch quantity and composition. Thus, in its simplest form, a catch-monitoring programme would be designed to document the quantity of fish harvested by a fishery and to provide this information with adequate temporal and spatial resolution. This type of data addresses information needs for scientists and managers. Managers must be satisfied that catch and bycatch limits are not exceeded and that catches are attributed to the correct gear types and management areas. Scientists must have accurate information on fishing mortality to conduct their stock assessments. They may also need information on size, age, sex ratio, maturity, feeding behaviour, and other biological characteristics by species, on the location of the catches, and on fishing effort. Scientists and managers may also need to monitor catches for lethal and sublethal bycatches of marine mammals and seabirds. Some fisheries are regulated heavily and may be subject to catch and bycatch limitations, vessel and/or gear-specific allocations, gear limitations, area closures, and seasonal restrictions. Under these circumstances, catches must also be monitored for compliance with fishing regulations.
Catches may be monitored at sea, during or immediately following gear retrieval, or on land, at the point of delivery. Catches may be monitored by industry personnel, or by independent observers who are trained to collect and report accurate information. For catches processed at sea, shore-based monitoring is not possible. Even when catches are delivered to land-based buyers and processors, they cannot be fully accounted for without at-sea observation. This is because discarding and high grading occur in most fisheries, even if such practices are prohibited. In recent years, there has been much institutional and public concern focused on global catch trends and the magnitude of wastage in fisheries, yet the data to document and investigate these concerns is often inadequate. Many of these data quality concerns can be resolved only through the use of at-sea observers.
Thus, while a comprehensive (or integrated) fisheries monitoring system may include vessel monitoring systems (VMS), paper or electronic logbooks, catch reporting by industry, and shoreside monitoring, it will, in many instances, require at-sea observation. Since observers are responsible only for gathering information, and they are expected to be independent and objective, they provide the only acceptable means for collecting certain types of fishery information. If vessel crew were assigned responsibilities for the collection and reporting of information, the results could easily become compromised by other vessel duties or conflicting interests. In addition, the standardised training offered by observer programme staff and the chain of custody provided for data, address the needs of scientists and managers for reliable and timely information. Observer programmes, however, are expensive to implement and maintain, and deployment of observers may be very difficult, particularly in small-scale, spatially dispersed fisheries. Furthermore, since monitoring data cannot be collected without error, accuracy and precision will often be of concern and this must be taken into account by those who design monitoring programmes and those who utilise the resultant data.
With these and other issues of concern to those who manage observer programmes and depend on the data they provide in mind, a Canada/U.S. observer programme workshop was convened at the Alaska Fisheries Science Centre in Seattle, Washington, USA in March 1998. The workshop was attended by about 80 people and included representatives of observer programmes in the United States and Canada, and monitoring programmes in Argentina, Norway and the Caribbean Community Region. Consistent with the broad range of issues faced by observer programmes and the diversity of stakeholders, participants were drawn from national and regional government agencies, non-government organisations, private contractors, observers, and fishing industry groups. Agency and private sector participants provided perspectives from a range of disciplines, including science, management, enforcement, policymaking and administration. Meeting participants represented a number of distinct observer programmes, with differing goals and objectives, and ranging in size from a few hundred to tens of thousands of observer days per year.
The perspectives which are offered on the use of observer programmes for monitoring commercial catches have evolved through our experiences with observer programmes in the groundfish fisheries of Alaska and British Columbia but have also been influenced by the discussions which took place at the workshop.
This review provides an overview of fishery observer programmes in the United States and Canada and also includes more comprehensive information about two specific programmes in Alaska and British Columbia. The broad range of factors which influence the quality of data collected by observers is discussed, including programme design and structure and the establishment of appropriate objectives. Recommendations which may be useful for organisations considering the establishment of at-sea observer programmes for catch monitoring purposes or evaluating the structure and performance of existing programmes are offered in conclusion.
Federal fishery observer programmes in the United States are implemented under the authority of three statutes: the Magnuson-Stevens Fisheries Conservation and Management Act, also referred to as the Sustainable Fisheries Act (SFA); the Marine Mammal Protection Act (MMPA); and the Endangered Species Act (ESA). While the conservation issues addressed by each of these acts differs, each requires data of the type which may only be provided by observer programmes, and two (SFA and MMPA) include specific provisions for fisheries observers.
Observer programmes authorised under SFA generally provide information on fish catch and bycatch, and may support information requirements for in-season management, compliance monitoring, and biological studies. Those authorised under MMPA and ESA are designed to evaluate fishery interactions with marine mammals, seabirds, sea turtles, and rarely occurring fish species, and may be focused on concerns involving threatened or endangered species. Some programmes may be implemented under the authority of more than one statute, with data collection objectives established accordingly. The U.S. programmes which were represented at the 1998 Canada/U.S. workshop are summarised below. While this summary does not provide information on all U.S. observer programmes, it does serve to document the types of programmes in operation, and some of their important details.
While most U.S. fishery observer programmes are implemented and managed by the federal government, some programmes are run by state government, and others by consortia which may include federal, state, and industry institutions. Sources of funding also vary. Most federal programmes are supported exclusively with government funds, but the largest federal programme, which monitors the groundfish fisheries off Alaska, has approximately 20% government funding with the balance provided by industry. The State of Alaska's shellfish observer programmes are funded in a similar manner, but some U.S. West Coast programmes receive partial funding from federal and state government, and industry. Overall, U.S. observer programmes provide approximately 60,000 - 70,000 coverage days per year.
A variety of service delivery models are employed. In some of the smaller federal programmes the observers are government employees, and in other programmes the government contracts directly for observer services. Observer services in support of the federal and state programmes in Alaska are provided by government-certified contractors. The agencies train and certify observers who are hired by these contractors. Fishing vessel and plant owners contract directly with observer companies for observer services.
In the North-eastern United States, observer programmes are operated by the National Marine Fisheries Service (NMFS), Northeast Fisheries Science Centre, in Woods Hole, Massachusetts, and the Commonwealth (State) of Massachusetts Division of Marine Fisheries. NMFS operates six distinct programmes in the Atlantic Sea Scallop Dredge Fishery, the Giant Atlantic Bluefin Tuna Purse Seine Fishery, the Large Pelagic Drift Gillnet Fishery, the Lobster Pot Fishery, the New England and Mid-Atlantic Gillnet Fisheries, and the Northwest Atlantic Trawl Fisheries. These programmes are designed to provide basic data for science and management, and they address a range of bycatch concerns and problems associated with marine mammal-fishery interactions. Small numbers of vessels participate in the tuna purse seine and pelagic long-line fisheries, and observer coverage levels are high. Large numbers of vessels participate in the other fisheries, and coverage levels are low. All programmes are mandatory except for the Northeast Atlantic Trawl Fishery, which has both mandatory and voluntary components. NMFS funds these programmes, and observer services are provided by a company which contracts directly to the agency.
The Commonwealth of Massachusetts maintains an observer programme which collects data in the lobster trap fishery and the trawl fisheries for squid, flounder, whiting, and shrimp. Coverage is mandatory in all fisheries but the level of observation is generally low. Programme responsibilities include; supplementing the NMFS sea sampling programme described above, providing information for development of fishing regulations, evaluating performance of experimental fisheries, and evaluating the consequences of such regulations as mesh size restrictions.
In the South-eastern United States, NMFS operates two observer programmes. The Swordfish Pelagic Long-line Observer Programme is operated by the NMFS Southeast Fisheries Science Centre in Miami, Florida. This programme is designed to collect data on effort, catch and bycatch quantity and biological characteristics, and information on fishery interactions with marine mammals, seabirds, and sea turtles. NMFS funds the programme and contracts direct with the individual observers. Between 2.5% and 5% of the effort is sampled, equivalent to approximately 4,000 observer days per year.
The South-eastern Shrimp Otter Trawl Fishery is operated by the NMFS laboratory in Galveston, Texas. The programme's mission is to characterise shrimp trawl bycatch and evaluate gear types designed to reduce bycatch, Interactions with sea turtles are of particular concern. Individual observers are contracted to perform this work. Coverage is voluntary, and less than 1% of the 250,000 fishing days per year in this fishery are monitored by observers.
In the Western United States (including Hawaii), NMFS operates four observer programmes. The Offshore Pacific Whiting Observer Programme is operated by the Alaska Fisheries Science Centre in Seattle, Washington. The remaining three, the California/Oregon Drift Gillnet Programme, the Hawaii Pelagic Long-line Programme, and the North Hawaiian Islands Lobster Programme are operated by the NMFS South-west Regional Office in Long Beach, California.
The Offshore Whiting programme involves 100%, voluntary coverage of a fleet of 16 trawl catcher/processors fishing off the U.S. West Coast; approximately 425 fishing days per year are observed and the data collected are used to support in-season management, catch and bycatch estimation requirements, and biological information needs for stock assessment. Since only 16 vessels are permitted to fish the offshore whiting quota, the companies involved have divided the quota into vessel-specific portions, and observer data is used by all participants to monitor the catches of each vessel and ensure that vessel-specific allocations are not exceeded. Observers are trained and certified by NMFS, and industry procures observer services from private, NMFS-certified observer contractors.
The mission of the California/Oregon Drift Gillnet Programme is to document the incidental take of marine mammals, sea turtles, seabirds, and target and non-target fish species, and collect biological specimens in the swordfish and thresher shark fishery. The programme is mandatory, and provides a coverage level of approximately 12% (421 observed sets in 1996). Costs are borne by NMFS, and observers are supplied by a private company which is under contract to the agency. In the Hawaii Long-line Fishery, NMFS deploys observers to document all species caught and collect biological specimens in this fishery which targets swordfish and several species of tuna. In addition to incidental takes of sea turtles, marine mammals, and seabirds, bycatch includes elasmobranchs and other large pelagics. NMFS funds the programme and employs the observers directly. In 1996, the six observers employed by this programme observed 5.3% of the fishing effort. The Northwest Hawaiian Islands Lobster Fishery Observer Programme is voluntary, and is designed to collect data on catch and bycatch quantity and composition, and to evaluate high-grading. In 1997, 66% of the 198 fishing days were monitored by observers employed by NMFS.
Three additional West Coast observer programmes were also discussed at the 1998 workshop. The West Coast Limited Entry Groundfish Trawl Programme and the Pacific Whiting Shoreside Landings Programme are both administered by the Oregon State Department of Fish and Wildlife (in association with several other agencies and industry organisations) and supported by industrial, regional and federal funds. The North Puget Sound Sockeye and Chum (salmon) Gillnet Fishery Observer Programme, was operated only in 1994 by the State of Washington Department of Fish and Wildlife. These programmes are all voluntary. The mission of the groundfish programme is to determine discard rates in these fisheries and to document the circumstances under which they occur. Coverage is approximately 5-10% and, in 1997, 245 vessels harvested the 98,000 metric ton (t) quota during a 125-day season. The Shoreside Landings programme is designed to document Pacific salmon and groundfish bycatch and collect biological samples required for stock assessment. The average annual catch is 61,000 t. In 1997, 14% of the landings were observed. Trawl fishery observers are employed by the State of Oregon in association with the Pacific States Marine Fisheries Commission. Observers participating in the shoreside landings programme are employed by state agencies or fishing companies.
The State of Washington salmon gillnet programme was designed to investigate the level of marbled murrelet and marine mammal entanglement in the non-treaty (non Native American) gillnet fishery in north Puget Sound. The programme was funded by NMFS and implemented by the State of Washington; 7% of sockeye fishing effort was observed during the 13-day season in 1994.
In the waters off Alaska, observer programmes are operated by the State of Alaska Department of Fish and Game (ADF&G) and the NMFS Alaska Fisheries Science Centre. ADF&G is responsible for managing shellfish in state and federal waters, and requires observer coverage in five specific fisheries; the Aleutian Islands King Crab Pot Fishery (all participating vessels observed), the Bering Sea Korean Horsehair Crab Pot Fishery (all participating vessels over 44 feet long observed), the Bering Sea King and Tanner Crab Pot Fisheries (all participating catcher/processors and processing vessels observed), and the state-wide Scallop Dredge Fishery (all participating vessels observed except in Cook Inlet). In all cases, ADF&G certifies observers who are trained by the University of Alaska, and fishing companies are required to procure observer services from private companies certified by ADF&G. In 1997, approximately 3,600 observer sampling days occurred in the crab fisheries, with an additional 500 days taking place in the scallop fishery.
Management of fisheries for groundfish in the U.S. Exclusive Economic Zone (EEZ) off Alaska began in 1977 following authorisation of the Magnuson Fisheries Conservation and Management Act (as it was then known), and the resultant regulations under which foreign fleets were allowed to harvest groundfish in U.S. waters. Initially, these fisheries were carried out exclusively by foreign vessels but, over a period of 15 years, they evolved through a joint venture phase and into the current situation where the quotas are harvested only by domestic entities. These fisheries are large and complex. Substantial quantities of groundfish are harvested in the EEZ off Alaska, with catches of approximately 200,000 - 250,000 t per year. being taken from the Gulf of Alaska, and 1.5 - 1.8 million t per year from the Bering Sea/Aleutian Islands region in recent years. Major species are walleye pollock (Theragra chalcogramma), Pacific cod (Gadus macrocephalus), yellowfin sole (Limanda aspera), rocksole (Lepidopsetta bilineata) and other flatfish, rockfish (Sebastes spp.), Atka mackerel (Pleurogrammus monopterygius), and sablefish (Anoplopoma fimbria). The fleet is large and diverse, and includes catcher vessels delivering to motherships and shoreside plants, and catcher/processor vessels which process catches at sea. Gear types include bottom and midwater trawls, long-lines, and pots.
Separate management plans have been established for the groundfish fisheries of the Gulf of Alaska and the Bering Sea/Aleutian Islands area and each of these plans establishes a quota harvest management system for the major groundfish species or species groups. Each year, stock assessments are performed, and the assessment scientists make recommendations for Allowable Biological Catches (ABCs). Based on these recommendations, the North Pacific Fishery Management Council recommends Total Allowable Catches (TACs) which are implemented by NMFS. TACs may be apportioned by area, season, and/or gear type and are generally managed in a manner which allows the major portion to be harvested by target-specific directed fisheries but provides sufficient quantities to meet bycatch requirements in fisheries where they are not targeted. Certain species which are taken in these fisheries may not be retained because their harvestable biomasses are fully utilised in other fisheries. These include Pacific halibut (Hippoglossus stenolepis), Pacific herring (Clupea harengus pallasi), salmonids, and several species of crab. In most cases, possession of these species on commercial groundfish vessels is illegal. However, this bycatch must be accounted for and, in the case of Pacific halibut, bycaught animals must be returned to the sea as quickly as possible and with minimal injury. Incentives to reduce bycatch of these Prohibited Species (PSC) include fleetwide bycatch quotas (or caps) in some fisheries, and requirements for individual vessels to keep certain PSC bycatch rates within limits in some fisheries (Pennoyer 1997).
In the Bering Sea/Aleutian Islands Area, a certain portion of the TAC (10% for pollock and 7.5% for other species) is set aside for the Community Development Quota (CDQ) programme. Under this programme, the CDQ is divided among a number of Western Alaska native community groups to be utilised for fisheries-related economic development. In general, these community groups have leased CDQ harvest rights to existing fishing companies so that these companies and, in some cases, the individual vessels which they own, may harvest only the specific quantity of CDQ catch that they have contracted to purchase.
The information requirements for managing these fisheries are substantial, and, since much of the catch is processed offshore, in-season management of quotas and PSC caps depends, to a large extent, on monitoring by independent fisheries observers. Furthermore, stock assessment scientists require detailed information on catch quantity, composition, and temporal and spatial distribution which would be difficult, or perhaps impossible to obtain without a large and well-trained corps of observers.
Placement of observers aboard foreign vessels harvesting groundfish off the U.S. West Coast and Alaska began in 1973. Initially, this was a small programme and observers were deployed only upon invitation of the host countries. The programme was greatly expanded in the late 1970s, when observer coverage on foreign vessels became mandatory, and a condition of the permits which were required of all participants in these fisheries, following implementation of the Magnuson Fisheries Conservation and Management Act (French et al. 1982). Coverage requirements were maintained as the fisheries evolved, from exclusively foreign, through a joint-venture phase to the exclusively domestic fishery, which has been in place since 1990. Regulations issued in 1990 established the current domestic observer programme with 30% coverage requirements for groundfish vessels of 60 ft - 125 ft length overall (LOA), 100% coverage requirements for vessels greater than 125 ft LOA, and coverage at shoreside processing plants based on the quantity of fish processed each month. In 1997, the programme deployed approximately 500 observers for about 30,000 coverage days.
The programme is expensive to operate since vessels are deployed from ports throughout Alaska and, in some cases, Washington and Oregon. The large factory trawlers may stay at sea for up to three months at a time, while many of the smaller catcher vessels must return to port every few days. The annual cost of operating the programme is $US10-12 million. Approximately $US2.5 million is provided each year by the federal government. This covers costs associated with training observers, providing them with safety and sampling equipment and supplies, supporting them in the field, debriefing them upon return from the field, and all the expenses associated with managing and disseminating the data they collect. Direct observer costs, including costs associated with deployment, observer salaries and benefits, and insurance are borne by the fishing industry. Under the service delivery model currently in place, NMFS is responsible for training and certifying all observers, and the observers are hired by one of the five private companies certified by NMFS to provide observer services. Fishing companies required to obtain observers must contract directly with one of these five companies. Observer companies charge between $US250 and $US300 per observer per day. In 1990, when this system was first put in place, it was recognised that this service delivery model was not consistent with the requirement for collecting independent, verifiable information and a commitment was made to devise a replacement procurement system which would be free from the potential for conflict of interest. This has yet to occur.
North Pacific groundfish observers are trained to perform a variety of functions including the collection of data on the quantity and composition of catch and bycatch, fishing effort, biological characteristics of catch species, fishery interactions with marine mammals and seabirds, and compliance with federal fisheries regulations. Observers file reports electronically on a weekly or daily basis, and their data are used in near real time by in-season managers who must project fishery closures based on their best estimate of when catch, bycatch, or PSC limits will be reached. Certain PSC species are taken in very small quantities, and observers must make special efforts to account for them and, in the case of Pacific halibut, observers must document handling procedures and the condition of halibut returned to the sea such that compliance with special handling regulations can be evaluated. Observers also monitor for compliance with a number of other regulations including those designed to reduce or eliminate discard of certain species, prevent the discharge of oil or dumping of general waste, and maintain gear type, time, or area based management restrictions. In the CDQ fisheries, the catch of each participating vessel is monitored by the observers on board, and fishing must cease when limits are reached. Observers monitoring in the long-line fisheries are trained to pay special attention to lethal bycatch of birds, and measures taken by crew to prevent them. Mortalities of the endangered short-tailed albatross (Diomedia albatrus) occasionally occur in this fleet, and can induce closure of the fishery if these mortalities exceed the small, allowable bycatch of this species. While uncommon, lethal bycatch of various marine mammals, some of which are threatened or endangered, also occurs.
Fishery observer programmes in Canada are authorised under the federal Fisheries Act for domestic fisheries and the Coastal Fisheries Protection Act (for foreign national vessels fishing within the Canadian EEZ). The Fishery (General) Regulations are the main instrument that provides for at sea observing activities by authorising the Department of Fisheries and Oceans (DFO) to place observers aboard fishing vessels, providing legal designation of observers, defining observer rights and duties, and defining the obligations of the vessel master toward observers. Information requirements of observer programmes are often mandated by fishery management plans and departmental policy, although the recently enacted Oceans Act will place a legislative requirement for certain types of fishery information available through observer programmes.
Observer programmes throughout Canada follow a consistent pattern of government control, programme delivery by private contractors and cost sharing by industry and government. DFO's Conservation and Protection Directorate provides national co-ordination for observer programmes, promoting programme consistency among the five regional DFO offices (four Atlantic and one Pacific) that each oversee and co-ordinate the delivery of individual programmes by contract with private sector firms. In general, there is a consistent approach to funding arrangements throughout Canada for domestic fisheries with DFO paying the programme administration costs (about 30% of the total programme cost) and industry paying for the costs of the observer. Foreign flag state vessels conducting fishing operations within the Canadian EEZ generally pay the full cost of observer programmes, including programme administration. The tri-party relationship between government, contractors and industry assigns contractors the responsibility for service delivery including: recruitment, training, deployment, briefing/debriefing and supervision of observers; data quality assurance and delivery; contract administration and management; and interaction with government and industry. Government is responsible for programme delivery through its control of the regulatory regime, the contracting process, requirements and administration, coverage levels, deployment strategy and policy, and product quality management. Industry is responsible for carrying observers as required and paying the associated costs.
The Government-contractor relationship is regulated through contracts and a Government observer-certification process. Observer service contracts are awarded through an open and competitive process, to establish a single observer service provider for each of the five DFO regions. By authorising a contractor the exclusive right to deliver observer services within a specific geographic area for a specific time period, this model avoids the conflicts of interest that would arise through direct service purchase negotiations between industry and the supplier. The arm's length relationship created through this service delivery model has been very effective in ensuring that programme outputs are both reliable and perceived as reliable by all stakeholders.
Certification is a prerequisite to working as an observer. Observer service providers hire, train and carryout certification examination for their observer candidates. Training generally takes three weeks to cover all aspects of the fishery, management and regulations, vessel operations and observer duties. In addition to meeting the training requirements, proficiency in at sea work must be met to retain certification. Recently, Canada has adopted national standards for observer training, certification, and code of conduct, providing more consistency between observers working on different programmes or regions (CGSB, 1997).
Canadian observer programmes began with the introduction of the 200 N.M. EEZ in 1977, as a measure by DFO to monitor activities of foreign fishing vessels while conducting operations within Canada's jurisdiction. As domestic fishing fleets expanded in capacity following the introduction of the EEZ, observer programmes were also introduced to these fleets and now have become part of the total management scheme for all major fisheries. Overall a corps of about 250 observers provides about 22,000 days of at sea monitoring per year within Canada.
In the Atlantic Region of Canada there are four separate observer programmes for each of DFO's administrative regions (Maritimes, Gulf, Newfoundland, and Laurentian). Close co-ordination between programmes is necessary as many fisheries and fleets span more than one region, resulting in observers from different Atlantic regions potentially contributing toward the monitoring of the same fishery. As a whole, the four observer programmes deploy about 180 observers for a total of 15,000 days per annum.
Atlantic Canada's large continental shelf area supports annual fish harvests of over 650,000 t (1996), principally shellfish (41%), pelagic fish species (39%), and groundfish (18%). Observer programmes play an important role in the management of these fisheries, ensuring their compliance with fisheries regulations and gathering vital fishery information that can not otherwise be obtained from fishing logbooks or shore-side landing inspections. The objectives of observer monitoring may vary by fishery but, in general, they are to monitor compliance with regulations (closed areas, small fish protocols, gear restrictions, high-grading, landing and reporting requirements) and gather data for stock assessment and to provide input into opening and closing fisheries.
Coverage levels for domestic fisheries are variable and usually between 5% and 20%. The northern shrimp fishery, which involves a fleet of 17 large catcher/processor vessels, operates with 100% observer coverage as do the majority of experimental fisheries and foreign vessels conducting fishing operations within the Canadian EEZ.
The North Atlantic Fishery Organisation (NAFO) Regulatory Area has been of interest to Canada because of potential impacts on fisheries within Canada's EEZ. Since 1996, NAFO Conservation and Enforcement Measures require observer monitoring of all NAFO member-nation vessels conducting fishing activities within the NAFO Regulatory Area. As a number of NAFO member countries did not have observer programmes when this regulation was implemented, Canada stepped forward to provide both funding and observers (through the Atlantic region's observer service contractors) to ensure the 100% coverage requirement was being met. The Canadian government also provided funds for training programmes, course curricula and other resource materials to assist the countries involved in developing their own observer service capability. Through these efforts, Canada supported about 3,800 observer coverage days in the NAFO region out of a total of approximately 12,000 observed days in 1997.
Within the Pacific Region of British Columbia, there are two main observed fisheries: the domestic groundfish trawl fishery and a foreign joint-venture trawl fishery that targets Pacific hake (also known as Pacific whiting). Observer monitoring occurs on a more limited basis in a variety of other fisheries including salmon (all gear types), shrimp trawl, groundfish hook and line fisheries, herring spawn-on-kelp, and experimental fisheries. In total, Pacific region observer programmes involve about 75 observers providing 6,000 data collection days per year.
Pacific Region ministerial policy raised compulsory observer coverage levels for the domestic groundfish trawl fleet to 100% in 1996. This decision recognised the critical need for providing of observers' independent, at-sea estimates for the management of the Individual Vessel Quota (IVQ) system, which was to be implemented in 1997. The fleet of 75 vessels fishes year round throughout offshore British Columbia with total landings of over 50,000 t. Over 80 fish species are captured in this fishery, and 22 of these are managed by area specific quotas. Thus, management of the fishery relies on accurate accounting of catch (including discards) for 55 different species-area quota units. Halibut bycatch and discards of unwanted fish are also significant management issues in this fishery. Bycatch rates are highly variable, with 10% of the total fishing operations accounting for nearly half the total quantities discarded and, conversely, nearly half the fishing operations result in about 10% of the total discarded catch. Thus, observer monitoring provides individual accountability for the fishing vessels, benefiting those with low bycatch rates and penalising those that catch and discard large quantities of unwanted fish.
Observers monitor all fishing operations at sea and all dockside-offloading operations to obtain accurate weights by species of landed fish. During a typical year, there are about 900 fishing trips or about 5,500 observed fishing days constituting approximately 20,000 trawl fishing sets. Fishery management utilises data collected at sea and dockside. The database updated upon the completion of each trip, in order to monitor overall catch and the quota holdings of each fishery participant. Biological sampling by dockside and at-sea observers meets an increasing amount of the scientific information needs for the stocks concerned.
In considering sampling and data quality, we first evaluate the factors which affect the accuracy of data collected by observers; these range from broad concerns regarding programme structure and appropriateness of objectives to more specific issues related to the difficulties associated with sampling on board commercial vessels, observer skill and experience, and the effects of inclement weather. Following the discussion of accuracy, some insights on sampling design are presented which demonstrate the extent to which sampling intensity influences precision when estimating catches of target and bycatch species.
Observer sampling is a three-stage process (Vølstad et al. 1997). The between-vessel level defines the proportion of the fleet observed, the within-vessel level is the proportion of hauls or sets within a vessel sampled by the observer, and the within-haul level concerns the selection of samples from individual catches.
Random selection of vessels is difficult and often impossible to achieve and is of concern only when coverage levels are less than 100%. When an observer programme works with a small, homogenous fleet that is not greatly dispersed in space and time, it may be possible for the programme manager to maintain familiarity with the activities of all participating vessels, and allocate observer coverage in a random manner. If vessel operators lack the incentive, however, to carry observers, they can readily subvert plans for random placement. Many programmes cover a diverse fleet that is spatially and temporally dispersed. In such situations, it is virtually impossible to track fleet operations and ensure the random placement of observers. Assumptions that the catch characteristics of observed and unobserved vessels within the same fleet are comparable may be difficult to support, particularly if compliance issues are involved.
Random selection of hauls or sets within a vessel is generally easier to achieve. In some cases, it may be possible for observers to sample all hauls or sets while aboard their vessels. In other cases, standard random sampling methodologies can be applied for selecting which hauls or sets will be sampled.
In some situations, it may be possible to select random samples of catch within a haul, or time intervals during the retrieval of a longline. An observer's ability to sample accurately, however, is influenced by many factors. These include: objectives and expectations, working conditions, sampling methods, and observer qualities. Because observer programmes are often designed to address multiple objectives and provide information to a range of clients, the integrity of their operation is of paramount concern. However, when reliable information is available from a fishery, different stakeholders, sometimes with opposing views, can focus on analysis and problem solving without becoming preoccupied with debates over the validity of the information.
Observer programme objectives may be narrow or broad. While there is a tendency to place many and varied responsibilities on the shoulders of observers, there is a finite limit to the amount of work they can accomplish, and requirements for training and previous experience may be greater for programmes with more demanding and complex sampling requirements. On the other hand, a programme designed simply to document specific types of fishery encounters with marine mammals or seabirds may underutilise its potential if observers are directed only to events which occur very rarely. Furthermore, when interactions with endangered species are concerned, it may be particularly important to document normal fishing activities and catch characteristics so that scientists can be provided with data to help them understand the circumstances under which these rare events occur.
Certain types of catch monitoring objectives can be addressed only by observer programmes. These include documentation of at-sea discards and high-grading, interactions with endangered or threatened species, collection of geo-referenced information on catch quantity, composition, and biological characteristics, and monitoring for compliance with certain types of fishery regulations (e.g. discard of oil and plastics, careful handling to reduce mortality of discards, vessel and haul specific bycatch rate limits), In other cases, observer data may be required to supplement or verify other sources of information such as for verification of logbook or catch reporting requirements, and monitoring for compliance with time, area, and/or gear restrictions.
Most programmes are designed to support multiple objectives, and, for many managers and policymakers, observer programmes are particularly valuable because of their ability to meet information needs for science, management, and enforcement. This makes it particularly important for those who design and manage observer programmes to recognise resource limitations, the feasibility of each objective, and the need to establish clear and unambiguous data collection priorities for observers. They must also understand that it may be necessary to compromise certain observer sampling objectives in order to ensure that other objectives are satisfied. It is also important to guard against placing unrealistic demands on these programmes and, hence, on the observers themselves. Over-ambitious tasking can lead to frustration when observer programmes cost more and deliver less than expected. Examples of conflicts which may arise as a result of incompatible objectives or over-ambitious tasking of observers include: placing unrealistically high scientific sampling requirements on a programme designed primarily to meet catch accounting and compliance monitoring goals in individual quota fisheries, and placing increasing demands for enumerating rare or protected species that must be returned to the water quickly, on a programme designed to provide accurate, haul-specific catch estimates. In these, and other similar situations, problems can sometimes be resolved by working with observers, industry, and other interested parties so that compromises can be reached without undermining the value of the data collected.
In addition to the need for identifying and prioritising programme goals, there is also a need for periodic reviews and adjustments to programme sampling designs and priorities which may also involve various stakeholders. Objectives and data collection priorities, however, are sometimes unduly influenced by political and economic considerations and under these circumstances conflicts may arise in the establishment of programme objectives or priorities. As a result, participants may fail to support the basic requirement for establishing programmes with achievable goals, and basic programme requirements, such as the need to provide observers with reasonable remuneration and suitable working conditions, may be compromised. Consequently, programme managers may find it difficult to maintain standards that encourage observer professionalism, and programme integrity may suffer. Evaluation of the overall performance of a programme should include review of its ability to respond to observer needs such as wages, resources, communication, programme involvement, safety and support.
Observer programmes are being used increasingly to monitor for compliance with fishery regulations. The deterrent effect of observers for most fisheries is a direct function of the coverage level. While there may be marginal compliance control value with low coverage level fisheries, high coverage levels are generally required to support this function. Because the presence of observers directly influences crew behaviour in these types of situations (and is intended to do so), data obtained from the observed portion of the fleet should not be used to make inferences regarding the unobserved portion.
In Canadian programmes, observers are directed to point out compliance concerns, giving fishers the opportunity to correct problems. This approach has resulted in more enforcement concerns being settled at sea, but with increased crew pressure on observers, particularly when the data they are collecting affects fishing or economic opportunity.
Observer-related cases are often difficult and time consuming to prosecute, and may be further complicated by the fact that, because there is a high turnover rate in some programmes, it may be hard to locate the observer once a case goes to court. The credibility of observer data is often an issue during the enforcement process, and defence lawyers may attempt to cast doubt on an observers' professionalism during prosecutions. Observer programmes that employ standardised protocols for recording and handling compliance related issues are, therefore, more successful in providing admissible evidence. Given these challenges, it is important for programmes to develop mechanisms to support and protect observers so that they can carry out their duties and, when necessary, be available to act as witnesses.
Observers may be reluctant to collect compliance-related data, particularly if they feel agency and contractor support is lacking. While fisheries enforcement personnel should prioritise actions to protect observers from interference or harassment, the personnel involved may have little understanding of observer responsibilities or the programme objectives. Observers may be reluctant to report compliance problems if they are unsure of the support which enforcement officials will provide. Programmes which allow fishing companies to negotiate freely with observer contractors may not be conducive to effective compliance monitoring by observers because fishing companies are free to terminate their association with a contractor and take their business to another "more co-operative" contractor if they feel it necessary.
Fishing industry groups have a vested interest in ensuring that independent, verifiable information is used to characterise fishing operations and evaluate the condition of the stocks. Under these circumstances, they can defend themselves against criticism from other industry groups and environmental organisations. Thus, from the industry's point of view, programmes must not only meet their objectives but must also be perceived by all clients to meet their objectives. Observer programmes often provide data to the fishing industry. This has, in some cases, supported activities directed at improving fishing gear and fishing strategies, increasing catch rates, decreasing bycatch, and developing new fisheries. These activities all tend to foster industry support of observer programmes.
The overall programme structure and service delivery model plays a significant part in creating an environment for observer professionalism and hence, providing high quality data. Historically, agencies have preferred to maintain in-house observer programmes with all functions being carried out by government personnel. In recent years, however, agencies have been under pressure to reduce employment levels and recover programme costs from industry. As a result, private sector participation in observer programmes has increased and this has provided increased flexibility with regards to meeting the often substantial observer programme labour requirements, work seasonality, and accommodating systems which allow for cost recovery from the fishing industry.
Observer programme managers are challenged with designing strategies to create an employment environment that attracts and retains high quality observers. Elements that contribute towards the development of professional staff include;
Regardless of whether service delivery is provided in-house (by government employees) or by private sector contractors, government retains the responsibilities for setting the rules under which an observer programmes exist. This may be accomplished by statute, regulation, and/or policy. Through this process, broad goals and objectives should be established, and requirements for service providers, fishing industry, and observer participants defined. Supplementary regulations and/or policies may implement management programmes which depend on monitoring by observers. Through these regulatory processes, agencies have an obligation to ensure that expectations are realistic and resources are adequate. They must also regulate to address issues regarding data collection integrity, safety, and suitable working conditions for observers. When compliance monitoring responsibilities are involved, or sampling interference and observer harassment are expected to occur, agencies are obliged to provide enforcement resources to back up observers and to ensure that the consequences of interference and harassment are appropriate.
Even when observer services are provided by contractors, it is generally preferable for agencies to retain responsibilities for such functions as training, certification, decertification, long-term data management and analysis. If all observer programme functions are contracted, agencies run the risk of losing programme functionality when a change in contractors occurs. This can be prevented by maintaining responsibilities for oversight and certain programme functions.
In some cases, the need to recover costs from the fishing industry results in the implementation of programmes with service delivery models under which fishing companies pay observer contractors directly for supplying observer services. This direct working relationship between contractors and industry may compromise government control and flexibility in placing observers, setting coverage levels, and designing sampling protocols. Furthermore, serious concerns about conflict of interest arise when companies compete for industry clients without government involvement and this process may erode confidence in the integrity of the data. Systems of this type also complicate and perhaps compromise government responsibilities for directing the placement of observers and holding contractors responsible for the quality of work performed by their employees.
The fishing industry is directly involved in the process of catch monitoring and must play a substantial role in the design and implementation of observer programmes if they are to be effective. Fishing companies are responsible for establishing a safe, effective, and respectful working environment for observers and assuring that crew co-operate with observers when appropriate. Unfortunately, certain types of data collection objectives may provide incentives for crewmembers to be unappreciative, suspicious, or even hostile. Under such circumstances, fishing companies and observer programmes must work together on creative solutions. Some programmes conduct observer placement orientations in order to familiarise skippers and crewmembers with observer duties, sampling protocols, and crew assistance requirements. Education and outreach efforts are more likely to succeed if there is a positive incentive for industry to co-operate. Noteworthy instances of co-operation have stemmed from working directly with fishermen and including them in research programmes.
The nature of the work setting aboard fishing vessels has a significant influence on the quality of information obtainable by observers. Fishing vessel design and operations are tailored to the efficient capture and handling of fish and involve processes which may not be conducive to data collection by observers. Very often, catch determination is a primary goal of observer programmes and observers direct the majority of their work time to catch monitoring tasks including estimating catch quantity and species composition, determining quantities of retained and discarded fish, and, in some instances, determining the viability of discards. Observers face a number of challenges in gathering this information; these may be caused by:
Basic observer sampling generally involves two major tasks: estimation of catch quantity, and collection of data for determination of catch composition. The most common include: visual estimation, basket sampling, counts applied to average weights, volume-density estimates, and whole haul catch enumeration. Even though visual estimation of catch quantity is discouraged due to the uncertainty and reliance on observer experience, it may, in some circumstances, be the only method available. Since weighing of catch at sea occurs in only a few fisheries, many programmes obtain catch weights by applying density coefficients to catch volume estimates. Several factors influence density, including the morphology of the species involved and the extent to which fish are compacted; since these factors may be difficult to take into account when obtaining density estimates, inaccuracies in catch quantity determination may arise (Dorn et al. 1997a). Most observer programmes strive to provide observers with a `tool kit' of sampling methods and guidance on practical approaches to use in particular circumstances. Programme managers depend upon observers to evaluate the overall circumstances of the work setting and employ the best methods for the particular situation.
The success of an observer programme is highly dependent on the qualities and strengths that the individual observers bring to it. Fishery observers are faced with unique challenges because their work occurs in remote, hostile locations, and sampling problems are often difficult to resolve. Nevertheless, many programmes have been successful because they have attracted high quality individuals and provided incentives for them to remain in the job. In contrast to a more centralised work setting, such as an office or factory or even a research vessel, where workers have regular and frequent supervisor contact, observer programmes are highly decentralised. Observers are isolated from each other and their supervisors and do not generally have the opportunity to seek advice when unique situations arise. Programme staff are generally unable to evaluate the observer methodologies and data until after the assignment has been completed. Since it may be difficult or impossible to monitor and correct problems during a trip, it becomes very important for managers to find ways to ensure that observer information is collected properly. Development and maintenance of a professional observer staff is, therefore, critical to the success of a programme and to the production of high quality data. The qualities which professional observers bring to a programme include:
Uncertainty associated with catch and bycatch estimates should be of concern to scientists responsible for stock assessments. However, catch estimation methodology does not always lend itself to the estimation of associated variances, and stock assessment methodologies generally consider catch information in a deterministic manner. In the Alaska groundfish fisheries, quota monitoring and, hence mortality estimation, is accomplished through an ad hoc process which uses observer data in some cases, industry reports in other cases, and stratifies according to the quantity and quality of data available in each temporal, spatial, gear type, and target fishery or bycatch category. Thus, it is difficult to link coverage levels directly with the information needs for the various objectives of the programme.
Nevertheless, since a substantial portion of the Alaska groundfish fishery is subject to 100% observer coverage, it is possible to develop catch and bycatch estimators with associated measures of uncertainty. A number of such studies have been carried out by NMFS at the Alaska Fisheries Science Centre, and they provide useful insights regarding coverage levels required to address different types of catch information requirements.
Dorn et al. (1997b) calculated the percent error (defined as: upper percent error = ((upper bound of the 95% confidence interval/estimate)-1) x 100, lower percent error = ((lower bound of the 95% confidence interval/estimate)-1) x 100) associated with estimates of catch and bycatch based on observer data in several target fisheries, although they considered only the variability associated with the proportion of vessels observed in a given fleet. Their analyses indicate that low levels of observer coverage are associated with relatively high levels of error for most target and non-target species (Fig. 1). For target species and some non-target species, error decreases rapidly as coverage increases from 10% to 30% and much more slowly under further coverage increases. For some bycatch species, however, uncertainty remains high, even when all vessels are observed.
Vølstad et al. (1997) conducted a detailed analysis of sampling and catch estimation in the (100% observed) trawl catcher-processor fisheries for walleye pollock and yellowfin sole in the Bering Sea/Aleutian Islands region. They developed a methodology for estimating catch and bycatch exclusively from observer data and provided associated measures of uncertainty. They also examined the trade-offs between uncertainty and the proportion of vessels in the fleet observed, and the proportion of hauls sampled by observers. In common with Dorn et al. (1997b), they did not consider uncertainty associated with within-haul sampling because appropriate data were not available. Also in common with Dorn et al. (1997b), they determined that the error associated with fleetwide estimates of target catch in these 100% observed fisheries was low, but that estimation uncertainty decreased for some of the major bycatch species, such as Pacific cod and rock sole (in the pollock target fishery), and much more so for the rarely occurring prohibited species, Pacific salmon and Pacific herring. The simulations they performed indicated that variability is high when either the fraction of vessels sampled within the fleet or the fraction of hauls sampled within a vessel is low. If random selection of vessels were possible, and the observer effect was not of concern, estimates of the catch of frequently occurring species with acceptable error bounds could be made by sampling approximately 30% of the vessels. For less frequently occurring species, a much larger proportion of the fleet would need to be sampled. Once aboard a vessel, observers generally sample 60-70% of the hauls, so the within vessel sampling fraction is generally not of concern. Examination of the results for bycatch of chinook salmon (Oncorhynchus tshawytscha) (Fig. 2) indicates the extent to which the precision of the overall chinook bycatch estimate is sensitive to the number of vessels observed and the fraction of hauls sampled within vessels, and that variability is much higher on some vessels than on others.
Even though salmon bycatch rates are very low in the pollock fisheries (substantially lower than 100,000 salmon in a total annual pollock harvest in excess of 1.0 million t in most years), certain highly productive pollock fishing areas may be closed when salmon bycatch limits are exceeded. The accuracy and precision requirements for salmon bycatch estimates, therefore, are high, but influenced by substantial sampling constraints. Since the work reported above considered only uncertainty associated with the between and within vessel sampling activities, Turnock and Karp (1997) conducted simulations to investigate uncertainty associated with the third sampling stage, the within haul level. Within haul sampling variability is of considerable concern when estimating bycatch of rarely occurring species which are likely to be highly aggregated. Since information on the within haul distribution is unknown, they made the simplifying assumption of a Poisson random distribution and simulated the sampling process to provide a basis for determining variability associated with each sampling stage. Because salmon bycatch may be limiting, and considerable between vessel differences are apparent, industry members are particularly keen to hold each fishing company responsible for its vessels' bycatch and fishing companies are keen to pass this responsibility on to their captains and crew. Consequently, managers are asked to provide vessel-specific salmon bycatch information, and they look to the observer programme for the necessary data. Recognising this need, Turnock and Karp (1997) paid particular attention to the sampling requirements for vessel specific sampling salmon bycatch determination. Their results suggest that variability decreases rapidly with increasing sampling intensity until the within haul sampling fraction approaches 20-30%, and that variability also remains high until 50-70% of the hauls are sampled (Fig. 3).
Figure 1. Percent error (defined as) as a function of observer (vessel) coverage associated with estimates of catch and bycatch species in the autumn 1996 trawl fishery for walleye pollock in the Bering Sea/Aleutian Islands region (from Dorn et al. 1997 b).
Fortunately, observers are generally able to sample 60-70% of the hauls taken by their vessels in the pollock fishery. Unfortunately, however, individual hauls often exceed 100 t, so the likelihood of sampling between 20% and 30% of the catch is small. Furthermore, the distribution of salmon in the catches is almost certainly more complex than was assumed during this analysis and, consequently, within-haul variability is likely to be much greater than predicted. Even though policy makers and managers may prefer individual vessel accountability for salmon bycatch in the pollock fisheries, practical limitations may preclude this possibility.
Figure 2. Coefficients of variation (CV) associated with chinook salmon bycatch estimates in the 1994 Bering Sea/Aleutian Island trawl fishery for walleye pollock. The upper figure illustrates the relationship between CV and the fraction of vessels and fraction of hauls samples. The lower figure illustrates vessel-specific CVs as a function of the fraction of hauls sampled for eight different vessels (from Vølstad et al. 1997).
Figure 3. Coefficients of variation associated with vessel-specific salmon bycatch rate estimates as a function of the within- and between haul sample fractions for the 1995 Bering Sea/Aleutian trawl fishery for walleye pollock. These results were obtained from a simulation based on the assumption of a Poisson random distribution of salmon are distributed within the pollock catches (from Turnock and Karp, 1997).
- Observer programmes provide the only means for collecting certain types of data required for determining the status of our living marine resources and determining the consequences of commercial fishery operations. Furthermore, they are an essential component of an integrated fisheries monitoring system;
- Since they are generally expensive and complex, great care should be taken when designing observer programmes to ensure that the goals and objectives are clear and unambiguous, and that they are achievable. Managers and scientists who depend on observer data must recognise the uncertainties involved, and place demands on programmes, which are consistent with these uncertainties. They must also recognise the consequences of placing conflicting demands on their observer programmes;
- To properly serve the needs of industry and government clients, the integrity of observer programmes should be carefully protected. This can be accomplished by ensuring that the service delivery model is free from conflict of interest, which is the case when observers are employed directly by government agencies or by contractors who are accountable directly to government agencies;
- Industry participation in the setting of programme goals and priorities should be encouraged, but should be protected from political influences which might result in placing unrealistic expectations on a programme. While industry co-operation is essential to successful programme operations, industry cannot be permitted to influence placement decisions made by programme managers. The benefits that industry obtains from observer programmes are greater when the public perceives that these programme are not subject to inappropriate industry influences;
- The conditions under which observers work are often harsh, and sampling requirements which are easy to achieve in a research environment may be impossible to achieve on a commercial vessel. While demands placed on observers must be realistic, the observers themselves must be provided with the tools necessary to carry out their duties. These include training and the provision of suitable sampling equipment, agency and contractor support, and maintenance of a safe, positive and productive working environment at sea;
- Observer experience and professionalism are essential to the success of any programme and steps should be taken to ensure that employment terms, working conditions, and all other aspects of programme design and implementation meet standards which will encourage individuals who perform well to observe on a continuing basis;
- The service delivery model employed by a programme may directly impact its ability to delivery high-quality, independent data. While certain programme functions should be retained by government agencies, provision of observer services by private companies who contract directly with government agencies is generally successful and cost effective. When fishing companies are permitted to negotiate directly with service providers, conflicts of interest may arise, and data integrity may be compromised. This concern is exacerbated when observers perform compliance monitoring duties;
- Agencies must retain sufficient expertise to allow them to transfer observer programmes to new contractors without interruption of service;
- When sampling interference or observer harassment occur, agencies are obliged to provide enforcement resources to back up observers and ensure that the consequences of this type behaviour are appropriate;
- Random placement of observers on vessels is generally difficult to achieve and when compliance concerns are involved it may not be reasonable to assume that the catch characteristics of observed and unobserved vessels are similar, and
- Even though it may be possible to select the hauls or sets to be sampled in a random fashion, requirements for random subsampling within a haul or set may also be difficult to meet.
CGSB, 1997. Training and Certification of At-Sea Fisheries Observers. National Standard of Canada CAN/CGSB-190.1-97. Canadian General Standards Board, Ottawa, K1A 1G6, Canada.
DORN, M., GAICHAS, S., FITZGERALD, S. & BIBB, S. 1997a. Evaluation of haul weight estimation procedures used by at-sea observers in pollock fisheries off Alaska. National Marine Fisheries Service, AFSC Processed Rep. 97-07: 76p. Alaska Fisheries Science Centre, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, Seattle WA 98115-0070, USA.
DORN, M., IANELLA, J. & GAICHAS, S. 1997b. Uncertainty in estimates of total catch for target and bycatch species at varying observer coverage levels in the Alaska groundfish fisheries. Unpublished manuscript.
FRENCH, R., NELSON, JR., R. & WALL, J. 1982. Role of the United States observer programme in management of foreign fisheries in the Northeast Pacific Ocean and the Eastern Bering Sea. North Am. Jo. Fish. Manag., 2:122-131.
PENNOYER, S. 1997. Bycatch management in Alaska groundfish fisheries. Am. Fish. Soc. Symp., 20 : 1414-1450.
TURNOCK, J, & KARP, W.A. 1997. Estimation of salmon bycatch in the 1995 pollock fishery in the Bering Sea/Aleutian Islands. Unpublished manuscript prepared for the North Pacific Fishery Management Council, Anchorage, Alaska, USA.
VØLSTAD, J.H., RICHKUS, W., GAURIN, S. and EASTON, R. 1997. Analytical and statistical review of procedures for collection and analysis of commercial fishery data used for management and assessment of groundfish stocks in the U.S. exclusive economic zone off Alaska. Project report prepared by Versar, Inc. for the National Marine Fisheries Service, (Available from William A. Karp, National Marine Fisheries Service, 7600 Sand Point Way NE, Seattle, WA 98115, USA).