THE PRECAUTIONARY APPROACH TO FISHERIES AND ITS IMPLICATIONS FOR FISHERY RESEARCH, TECHNOLOGY AND MANAGEMENT: AND UPDATED REVIEW

by S.M. Garcia

Fishery Resources Division, FAO Fisheries Department

Via delle Terme di Caracalla, 00100-Rome, Italy

INTRODUCTION

There is an obvious link between the sustainable development of fisheries and their precautionary management. In 1988, the 94th Session of the FAO Council agreed that “Sustainable development is the management and conservation of the natural resource base, and the orientation of technological and institutional change in such a manner as to ensure the attainment and continued satisfaction of human needs for present and future generations. Such development conserves land, water, plant genetic resources, is environmentally non-degrading, technologically appropriate, economically viable and socially acceptable.” This definition applies well to sustainable fisheries development and management.

The strategies required to ensure a high degree of sustainability in human use of natural renewable resources systems are not easy to conceive and implement for at least two reasons: (a) our insufficient understanding of the laws governing these systems and the inherent uncertainty about the consequences of our decisions, and (b) the inadequate nature of our institutions and controls (Holling, 1982; 1994), particularly on access to resources. It is generally agreed that the inadequacy in management results essentially from the open access nature of the fisheries and the lack of effective mechanisms to directly control fishing effort levels in the absence of an explicit agreement on the allocation of resources between users. It is also being realized that, in addition, the problem lies partly in the non-recognition of the high levels of uncertainty that characterize fisheries and the related lack of precaution in most management regimes. The review of the state of world fishery resources undertaken by FAO and the global analysis available in the FAO report on the State of Food and Agriculture (SOFA) show that, although management practice has favourably evolved during the last half century, it has tended to lag behind management theory and that progress towards sustainability, since the first FAO Technical Committee on Fisheries in 1945, has been insufficient. It is now recognized that the biomass of many important fish stocks is close to or even below the level that could produce the maximum sustainable yield (MSY), leading to resource instability and economic losses. A number of fisheries have collapsed ecologically or economically and the situation in the high seas raises particular concern. In many areas, the present situation is one of resource erosion, economic losses and social dislocations that illustrate the fisheries management risk and reflect behaviour which in the last decades has been neither sufficiently responsible nor precautionary (Garcia, 1992; FAO, 1993; Garcia and Newton, 1994; 1995).

The increased recognition that conventional fishery management needed to be improved has been accompanied by a growing concern for environmental management, particularly as a result of the World Conference on Human Environment (Stockholm, 1972), the FAO Technical Conference on Fishery Development and Management (Vancouver, 1973), the FAO World Conference on Fisheries Management and Development (Rome, 1984), the United Nations Convention on the Law of the Sea (hereafter, the 1982 Convention), the work of the Brundtland Commission from 1984 to 1987 (World Commission on Environment and Development, 1987), the United Nations Conference on Environment and Development (Rio de Janeiro, 1992), the International Conference on Responsible Fishing (Cancun, Mexico, 1992) and the UN Conference on Straddling Fish Stocks and Highly Migratory Fish Stocks (New York, USA, 1993–1995). Moreover, the emerging awareness of the complexity of marine ecosystems and related scientific uncertainty, particularly in the high seas, and of the risk of error in management, requires an acceleration of the evolution of fishery management, a broadening of its scope and a change in attitudes. Two important and related requirements of the new management context are the need for more caution and for better inter-generational equity. The latter issue concerns the ethics of renewable resource use and the moral obligation placed on the current generation to exploit the resources and enact conservation measures in such a manner as to preserve options for future generations.

The poor control of fisheries development by fishery management authorities is one of the major reasons for the present state of fisheries. In natural ecosystems, the abundance of preys and predators, and their variations, are controlled and maintained within limits compatible with the ecosystems sustainability by a set of complex interactions and feed-back mechanisms. In ecological terms, fisheries are organized “top predators”. As such, their survival depends on the survival of their living resources and they are certainly far more sensitive to natural feedback information on the state of the resources they exploit than industrial systems using oceans as a resource for waste-dumping. However, contrary to natural predators, fishermen are not entirely controlled by feedback signals of resource stress. Their operations are not totally dependent on the abundance of the various elements of the resource ecosystem and, indeed, are partly isolated from such feedback controls by various mechanisms such as price increases (as resources become scarcer), technological improvements in efficiency, shifts to other species or areas, and governmental subsidies. They can, therefore, continue and even expand their operations despite the environmental and resource degradation they may produce.

Section 1 of the document defines the objectives of the precautionary approach in the specific field of fisheries. Section 2 proposes some definitions of key concepts used in the document. Section 3 provides an updated review of trends and perspectives in the development in the concepts and applications of the principle of precautionary action, including both the precautionary principle and precautionary approach. Section 4 concentrates on one of the major issues related to, and indeed justifying, precaution such as the uncertainty due to incomplete knowledge, the potential errors in decision-making and the consequent potential risk. Sections 5, 6 and 7 describe the implications of the precautionary approach and provide practical guidance for its application in the respective areas of research, technology development and transfer, and conservation and management. The conclusion provides a summary of the approach and its prospects, focusing particularly on management.

1. OBJECTIVES OF THE PRECAUTIONARY APPROACH

The modern requirement to deal explicitly with uncertainty, in order to reduce risks to the resources and their environment (and indeed to the fishing communities), requires significant changes in the fields of science, technology and fishery management. Such changes are required in order to effectively deal with the unprecedented shift in policy and international relations and with the metamorphosis of public perceptions and political demands resulting from the 1982 UN Law of the Sea Convention, UNCED and its Agenda 21. One of the elements of change is the requirement for a more precautionary approach to natural resources management. The concept of precautionary action aims generally at improving conservation of the environment and the resources by reducing the risk of inadvertently damaging them. More specifically, it aims at helping decision-makers and regulators to take a safeguarding decision, when the scientific work is inconclusive but a course of action has to be chosen. In addition, it intends to promote a more equitable balance between the short-term considerations (which led to the present environmental degradation and overfishing) and long-term considerations such as the need to conserve resources for future generations. It aims at promoting inter-generational equity by reducing the cost of our decisions for future generations and by counteracting the effects of current high economic discount rates which provide a strong incentive to overfish, maximizing the discounted net benefits from a stock and, de facto, giving preference to present consumption over future consumption¹. By comparison, and despite the fact that it theoretically aims at sustainability, conventional fishery management addresses primarily, and rather inefficiently, the issue of inter-generational equity and allocation of resources between present users. The concept of precautionary action will also directly benefit present generations of fishers and consumers if fishery authorities and industry actively promote its implementation by other economic sectors whose activities damage ocean productivity, fishing communities' livelihood and consumers' health².

¹This factor often leads to proposals to introduce a social discount rate. However, there are severe practical difficulties in determining and implementing such rates. A more satisfactory solution would appear to be through proper pricing of resources, including not only the marginal cost of harvesting, but also the foregone value of catches no longer available to future generations

²Opportunity to promote this approach is given by the growing requirement to integrate coastal fisheries management into the Integrated Coastal Areas Management (ICAM) within which inter-sectoral competition for resources should be organized and controlled

2. DEFINITIONS

The literature on the precautionary principle or approach is loaded with terms the meaning of which may not always be obvious or universally agreed and, in order to facilitate common understanding, this section proposes some definitions with their source. The original ones draw heavily from the discussions in the following sections and should be considered together with them.

Acceptable impact: A negative, or potentially negative, alteration of the exploited natural system, resulting from human activities (i.e., fisheries and other impacting industries), the level and nature of which, on the basis of available knowledge, is considered as representing a low enough risk for the resource, system productivity, or biodiversity. Its acceptability is continuously kept under review and can be revocated on the basis of new knowledge.

Approach: “A way and means of reaching something. The method used in dealing with or accomplishing something” (Houghton Miflin Co., 1992).

Precaution: “An action taken in advance to protect against possible danger or failure; a safeguard. Caution practised in advance. Forethought or circumspection” (Houghton Miflin, 1992). Action taken in advance of scientific certainty but within the bounds of scientific uncertainty, to avoid or minimize negative impact, taking into account the potential consequences of being wrong (modified from a definition in relation to global climate change by Turner, O'Riordan and Kemp, 1991).

Precautionary approach: A set of agreed cost-effective measures and actions, including future courses of action, which ensures prudent foresight, reduces or avoids risk to the resources, the environment, and the people, to the extent possible, taking explicitly into account existing uncertainties and the potential consequences of being wrong³.

Principle: “A basic truth, an assumption. A rule or standard, especially of good behaviour. A fixed or predetermined policy or mode of action” (Houghton Miflin, 1992)⁴.

Reference points: “A (management) reference point is an estimated value derived from an agreed scientific procedure and an agreed model to which corresponds a state of the resource and of the fishery and which can be used as a guide for fisheries management”⁵:

Limit Reference Point (LRP): indicates the state of a fishery and/or a resource which is not considered desirable. Fishery development should be stopped before reaching it. If a LRP is inadvertently reached, management action should severely curtail or stop fishery development, as appropriate, and corrective action should be taken. Stock rehabilitation programmes should consider an LRP as a very minimum rebuilding target to be reached before the rebuilding measures are relaxed or the fishery is re-opened.

Target Reference Point (TRP): corresponds to the state of a fishery and/or a resource which is considered desirable. Management action, whether during a fishery development or stock rebuilding process, should aim at maintaining the fishery system at its level.

Threshold Reference Point (ThRP): indicates that the state of a fishery and/or a resource is approaching a TRP or a LRP, and a which a certain type of action (usually agreed beforehand) needs to be taken. Fairly similar to LRPs in their utility, the ThRPs' specific purpose is to provide an early warning, reducing further the risk that the TRP or LRP is inadvertently passed due to uncertainty in the available information or to the inertia of the management and industry system. Adding precaution to the management set-up, they might be necessary only for resources or situations involving particularly high risk.

Risk: In general, “the possibility of suffering harm or loss; danger. A factor, thing, element, or course involving uncertain danger, a hazard” (Houghton Miflin, 1992). In decision theory “the degree of probability of loss. A statistical measure representing an average amount of opportunity loss” (Kohler, Cooper and Ijiri, 1983). This terminology is used “when large amounts of information are available on which to base estimates of likelihood, so that accurate statistical probabilities can be formulated” (Pass et al., 1991). The Technical Consultation on the Precautionary Approach to Capture Fisheries (FAO, 1995), in this case, refers instead to “expected loss” or “average forecasted loss” to clearly distinguish between the general meaning and the decision-theoretic one (see also Shotton, 1993).

Risk analysis: “Any analysis of unknown chance events for purposes of effecting or evaluating decisions in terms of possible penalties and benefits attending these events. A method for generating different probability distributions with accompanying cost and benefits that may attend different courses of action. Generally uses computer simultations” (Kohler, Cooper and Ijiri, 1983).

Uncertainty: “The condition of being uncertain. Doubt. Someting uncertain. In statistics, the estimated amount or percentage by which an observed or calculated value may differ from the true value” (Houghton Miflin, 1992). “The incompleteness of knowledge about the states or processes in nature” (FAO, 1995).

³There is paradoxically no definition of the precautionary approach which is generally related to the need to take action even in the absence of “full scientific certainty” and defined by its implications. This definition has been developed by the author based on the definitions of “precaution” and “approach”, above, and on UNCED Principle 15

⁴It can be noted that while the first part of this definition differentiates between the precautionary “principle” and “approach”, the second part tends to blur the difference between the two concepts

⁵According to the ad hoc Working Group on Reference Points established by the UN Conference on Straddling Fish Stocks and Highly Migratory Fish Stocks in New York, in March 1994 (cf. Annex 5)

3. TRENDS AND PERSPECTIVES

There is no explicit reference to the principle in the 1982 Convention. Part XII, on “Protection and preservation of the marine environment”, does not contain detailed instruments for implementation of the conservation of the marine ecosystem, but it does state in a global instrument, in article 192, the following general obligation: “States have the obligation to protect and preserve the marine environment” (Burke, 1991). In addition, ecosystem conservation also requires measures for the fisheries sector, striking a balance between the provisions for environmental conservation and fisheries management to ensure sustainable exploitation.

However, in fisheries, the concept of precautionary action seem to have progressively become an important factor in negotiations between States to establish management measures in circumstances where there is an obligation to negotiate in good faith to reach agreement (e.g., with respect to highly migratory, straddiing or shared fish stocks, under the 1982 Convention). It can be assumed that, given the wide support for this concept in environmental law, a State which refers objectively to it will hope that it cannot be accused of bad faith (Burke, 1991). The concept is also developing in national fisheries management regimes. The concept of precaution has been expressed as “the precautionary principle” (hereafter, the principle) or “the precautionary approach” (hereafter, the approach). Although the two terms relate equally well to the concept of caution in management, and sometimes not differentiated by scholars (e.g., Bodansky (1991) uses the two terms alternatively), they are differently perceived by international lawyers, negotiators and industry, as shown below. The term “approach” is apparently more generally accepted by Governments in the fisheries arena because it implies more flexibility, admitting the possibility of adapting technology and measures to socio-economic conditions, consistent with the requirement for sustainability. It is particularly more appropriate for fisheries because consequences of errors in their development or mismanagement are unlikely to threaten the future of humanity and, in most cases, are reversible. On the contrary, the term “principle” has developed a negative undertone because it is usually given a radical interpretation and has led to the outright ban of technologies, e.g., in the case of whaling (Bodansky, 1991) and the Large Scale Pelagic Driftnet Fishing (see below), and is sometimes considered incompatible with the concept of sustainable use. These two concepts are further elaborated below.

3.1 The Precautionary Principle

This principle's most characteristic attributes are that: (a) it requires authorities to take preventive action when there is a risk of severe and irreversible damage to human beings; (b) action is required even in the absence of certainty about the damage and without having to wait for full scientific proof of the cause-effect relationship, and (c) when there is disagreement on the need to take action, the burden of providing the proof is reversed and placed on those who contend that the activity has or will have no impact.

It seems generally agreed that the precautionary principle has originated in Germany as the “Vorsorgenprinzip” (Dethlefsen et al. 1993). The principle has been referred to and applied at national level in relation to human activities with potentially severe effects on human health (engineering, the pharmaceutical and chemical industries, nuclear power plants, etc.). In international environmental law, the principle has emerged as a recognition of: (a) the uncertainty involved in measuring the impact of toxic substance on the ecosystem and the human health, and (b) deciding on the “assimilative capacity” of such ecosystems (i.e., their ability to absorb a certain quantity of the substance in question without unacceptable impacts). In the 1970s, following the 1972 Stockholm Conference, concern for human safety was progressively extended to the human environment and to other species. This led to increasingly frequent reference to the principle in international agreements and conventions, often with limited consideration of its practical implications. It has been introduced at international level at the First International Conference on the Protection of the North Sea (1984) in relation to persistent toxic substances susceptible to bioaccumulation in the marine ecosystem. The 1987 Declaration of this Conference contains an example of the concept of precaution in relation to coastal States' jurisdiction, habitats, species and fisheries, including pollution from ships. It provides that “States accept the principle of safeguarding the marine ecosystem by reducing dangerous substances, by the use of the best technology available and other appropriate measures” and that “this applies especially when there is reason to assume that certain damage or harmful effects on the living resources are likely to be caused by such substances and technologies, even where there is no scientific evidence to prove a causal link between practices and effects.”

The scope of application of the precautionary principle was successively broadened from persistent toxic substances to all synthetic persistent substances, natural substances released in large quantities (e.g., nutrients responsible for eutrophication) and finally to all emissions responsible for global warming (Dethlefsen et al., 1993). The principle has been invoked in issues related to the ozone layer (1985 Vienna Convention for the Protection of the Ozone Layer and the 1987 Montreal Protocol on Substances that Deplete the Ozone Layer) where States agreed to reduce emissions of certain substance at a time when the causal links had not yet been firmly established (Boelaert-Suominen and Cullinan, 1994). It has also ben referred to in relation to the greenhouse effect and the conservation of nature. It has touched indirectly on fisheries through provisions in the international convention on dumping at sea (the Paris and Oslo Conventions, Marpol) relating to pollution by fishing vessels. The 1991 International Conference on an Agenda of Science for Environment and Development into the 21st Century (ASCEND 21) referred to the principle, stressing “the central importance of the precautionary principle according to which any disturbance of an inadequately understood system as complex as the Earth system should be avoided”. Broadus (1992) asked whether that meant “any disturbance” and at “any cost” indicating that the principle was not a principle but a range of more-or-less rhetorical prescriptions for choice in front of uncertainty. The principle has also been considered as particularly appropriate in the context of Integrated Coastal Areas Management (Boelaert-Suominen and Cullinan, 1994) because of the vulnerability of coastal resources, the likelihood of swift and irreparable harm, and the incomplete understanding available on the complex web of interconnected biological processes in the coastal area. More recently, the precautionary principle has also implicitly been included in the Convention on Biological Diversity (UNEP, 1992) which noted, in its preamble “that, where there is a threat of significant reduction or loss of biological diversity, lack of full scientific certainty should not be used as a reason for postponing measures to avoid or minimizing such a threat.”

The principle remains contentious both within the scientific community and from the point of view of policy-makers and these controversies are illustrated in the fact that there is, as yet, no generally accepted formulation of the principle. When the interpretation of the principle is softened, the border between it and the approach is significantly blurred. For instance, Young (1993, cited by Dovers and Handmer, 1995), proposes to consider four different levels of application of the principle, corresponding to decreasing levels of risk, potential degree of irreversibility, and uncertainty:

Level 1: Impacts are potentially serious (unacceptable) or irreversible and uncertainty is high: a strict application of the principle is required, insisting on complete reversibility and putting a strong burden of proof⁶ on development proponents.

Level 2: Impacts may be serious but potentially reversible and a reasonable amount of data is available to appreciate risk: large safety margins should be ensured in assessments and decisions and use of the best available technology should be strictly required, i.e., regardless of costs.

Level 3: Impacts are considered largely acceptable (and/or potentially reversible) and reasonably good scientific and other information is available: lower safety margins are accepted. The best available technology is required only if economical.

Level 4: Potential losses are considered neither serious nor irreversible: decisions could be based on traditional cost-benefit analysis.

⁶See discussion on the burden of proof in Section 5

The conditions for the application of levels 3 and 4 and their implications are very similar to the conditions and implications of the precautionary approach and illustrates that these two related concepts are sometimes difficult to distinguish.

The large-scale pelagic driftnet issue

The UN General Assembly Resolution 44/225 of 22 December 1989, on large-scale pelagic driftnet fishing and its impact on the living marine resources of the world's oceans and seas, could be considered a case of radical application of the concept of precaution, despite the lack of explicit reference to the principle. The resolution expressed concern about the size of the fleets, the length of the nets, their mode of operation, their potential impact on anadromous and highly migratory species, their by-catch and the concern of coastal countries on the state of resources close to their exclusive economic zones. It recommended that a worldwide moratorium should be imposed on all driftnet fishing by 30 June 1992 and it established a set of immediate and regionally tailored interim measures. It also provided that such measures would not be imposed in a region or, if implemented, could be lifted, should effective conservation and management measures be taken upon statistically sound analysis to be made jointly by concerned parties. The proposal is rational but the flaws in the process followed for the implementation of the resolution have been underlined (Miles, 1992, 1993; Burke, Freeberg and Miles, 1993).

The consequences of this resolution, after heated international debate and political pressure, has led to the discontinuation of the issuance of fishing licences and research for alternative fishing techniques, in Japan and Taiwan (Province of China); the docking and conversion of driftnet fishing vessels in the Republic of Korea and a regulation by the European Union (see below). Large-scale driftnet fishing stopped in the South Pacific in 1992–93 but some fishing continued in the Mediterranean and Bay of Biscay, where scientific experiments were conducted to assess the fishery's impact on the associated small cetaceans. Many other Mediterranean countries, however, have taken regulations prohibiting driftnet fishing in their waters. Following up on the UN Resolution, the European Community adopted a Council Regulation (N° 345/92 of 27/1/1992) limiting to 2.5 kilometres the length of the driftnets authorized, but granting a derogation to 5.00 kilometres, until 31 December 1993, to vessels having fished for at least three years preceding the implementation of the regulation. This derogation was to expire by the indicated date unless scientific evidence showed the absence of “any ecological risk”.

3.2 The Precautionary Approach

In considering the introduction of more precaution in fisheries management and development, the main differences between fisheries impacts and chemical industries pollution (for the control of which the precautionary principle was created) must be kept in mind:

1. the assimilative capacity in relation to fisheries impact (i.e., the quantities of fish that can be removed without damaging the system's productivity) exists without doubt and can be determined with some accuracy, even though it varies, and

2. the impacts are, in most cases, reversible and, as a result, the potential consequences of an error would rarely be dramatic, even though they can be significant in socio-economic terms.

In the early 1990s, the precautionary approach has been progressively more accepted and its field of application has been broadened to include the management of natural renewable resources, including fisheries. The aims of the precautionary approach are similar to those of the precautionary principle from which the approach is sometimes difficult to distinguish. The main difference between the principle and the approach might be that the latter considers explicitly the social and economic implications of its application in order to ensure that: (a) it does not lead to imbalance in favour of non-fishery uses and future generations with undue strain on present generations and the fishery sector, and (b) that unavoidable short-term costs to the fishery sector are mitigated and equitably shared. The various interlinked processes that lead to the widespread adoption of the precautionary approach in fisheries, are briefly described below.

The UNCED process

UNCED stressed the need for a precautionary approach to ocean development in its Rio Declaration and in Agenda 21, particularly in its chapters on the management of coastal areas, resources under national jurisdiction and high seas resources. The principle 15 of the Declaration states that “in order to protect the environment, the precautionary approach shall be widely applied by States according to their capabilities. Where there are threats of serious or irreversible damage, lack of full scientific certainty shall be not used as a reason for postponing cost-effective measures to prevent environmental degradation.” The wording, largely similar to that of the principle, is subtly different in that: (1) it recognizes that there may be differences in local capabilities to apply the approach, and (2) it calls for cost-effectiveness in applying the approach, e.g., taking economic and social costs into account. UNCED led to agreement on two principles which are intuitively reasonable and potentially contradictory: the precautionary approach and the principle of economic efficiency. The delicate co-existence of these two principles impedes the development of safeguards against uncontrolled decisions (or lack of decisions).

The FAO process

Many years before the issue became fashionable in the fisheries circles, FAO, through its European Inland Fisheries Advisory Commission (EIFAC), collaborated with the International Council for the Exploration of the Sea (ICES) in the development of ICES/EIFAC Codes of Practice and Manual of Procedures for Consideration of Introduction and Transfer of Marine and Freshwater Organisms (Turner, 1988)⁷. This Code stresses that, in a context of rapidly changing population pressures, the impact of the introduction of species to enhance the potential of sustainable fisheries should be examined in the light of the likely impacts of alternative development strategies, involving environmental degradation and likely to result in changes in species composition of both the terrestrial and aquatic ecosystems.

More recently, in a review of the FAO programme in marine fisheries management, Garcia (1992) identified some of the challenges to be faced by fisheries in the period 1993–2000. These included: the uncertainty in the scientific information, the need for a more precautionary approach to management, the burden of proof and the need to define “acceptable” levels of impact. At the 1992 FAO Technical Consultation on High Seas Fishing, Garcia (1992a) stressed the uncertainty in the “best scientific evidence available” for management and drew attention to issues of precaution and burden of proof, the non-precautionary nature of the traditional MSY reference point, and the need for more and different reference points to be used as a basis for more precautionary management strategies. The Consultation provided guidance to the Fisheries Department of FAO on how to proceed (FAO, 1992) and, inter alia, agreed that:

• fisheries should be managed in a cautious manner;
• precaution did not necessarily require a moratorium on fishing;
• there was a need to identify methods to handle uncertainties;
• the objective was to safeguard both people's livelihood and biodiversity;
• existing precautionary measures should be included in the Code of Conduct;
• precautionary measures should be based on science and not be discriminatory, and
• measures should be revised or revoked when new information became available.

⁷ A full-scale practical application of this Code has been undertaken by FAO in Papua New Guinea (Coates, 1994). starting from the premise that introductions of new species in an aquatic ecosystem should be subject to prior evaluation. irrespective of whether species are “exotic” or not

The International Conference on Responsible Fishing (Mexico, 6–8 May 1992), organized in close cooperation with FAO, defined the concept of responsible fishing as encompassing “the sustainable utilization of fishery resources in harmony with the environment; the use of capture and aquaculture practices which are not harmful to ecosystems, resources or their quality; the incorporation of added valued to such products through transformation processes meeting the required sanitary standards; the conduct of commercial practices so as to provide consumers access to good quality products”. The Cancun Declaration contains a fairly complete prescription for modern fishery management covering environmental impacts; multispecies by-catch and discards issues; effort control requirements; etc., but did not include any explicit reference to the precautionary approach. One year later, however, the Inter-American Conference on Responsible Fishing (Mexico City, July 1993) referred to the need to take precaution into account in the Code of Conduct on Responsible Fishing, particularly in the high seas.

In 1993, the review of the state of highly migratory species and straddling stocks, prepared by FAO at the request of the UN Conference on Straddling Fish Stocks and Highly Migratory Fish Stocks indicated that it was necessary “to analyse the potential role and agree on possible ways of implementing cautious management approaches compatible with sustainable fisheries” (FAO, 1994, page 65). Following a first attempt to analyse in detail the various implications of the concept of precautionary action in fisheries research, management and development (Garcia, 1994), a document was prepared by FAO, to comply with a request by the UN Conference on Straddling Fish Stocks and Highly Migratory Fish Stocks (Second Session, July 1993). This document (United Nations, 1994; Garcia, 1994a) was presented to the UN Conference at its meeting of March 1994. Even though it was prepared for a meeting on straddling and highly migratory resources, the document was considered by FAO as generally pertinent for all resources and fisheries, whether in the high seas or under national jurisdiction, because it was felt and stated that, if a resource required precaution, it should be provided regardless of the type of jurisdiction, and the set of management measures applied to the various life stages of a transboundary resource should be coherent across its entire area of distribution. Unfortunately, this logical and basic biological requirement became, at the UN Conference, one of the major points of disagreement because some coastal countries considered that the need for overall “coherence” or compatibility between the management regimes inside and outside the EEZ could represent or be interpreted as an encroachment on their sovereign rights⁸.

The issues of scientific uncertainty and precaution were also addressed in another document prepared by FAO for the UN Conference on Straddling Fish Stocks and Highly Migratory Fish Stocks, on management reference points (United Nations, 1994a; FAO, 1994). This report recognized that “most of the difficulties experienced in using any target reference point results from the considerable uncertainties as to the current position of the fishery in relation to it”. It suggested using limit reference points (LRPs) as a way to increase the precautionary nature of the management set-up. Such LRPs, to be used alone or in combination, could correspond, for example, to situations where: (a) spawning biomass or proportion of mature individuals fall below, say, 20% of the values for the virgin stock; (b) fishing mortality falls below, say, 30% of the virgin stock biomass-per-recruit or reaches 80% of the rate of natural mortality; (c) total mortality reaches the level corresponding to Maximum Biological Production for the stock; (d) mean individual size fall below the mean size at maturity; (e) annual recruitment levels remain below a certain level (or average level) for a certain number of years, and (f) the resources rent have been totally dissipated (i.e., the total cost of fishing, including reasonable revenues to manpower and capital, are equal to total revenues), etc.

⁸A situation could be foreseen in which a sovereign coastal State could see its right to introduce a technology (e.g., a new fishing gear, or practice, or genetically modified organisms) questioned by non coastal countries exploiting the same straddling or highly migratory stock

FAO has started the preparation of a Code of Conduct for Responsible Fisheries following the International Conference on Responsible Fishing, held in Cancun (Mexico, 1992). The Code includes a section on precautionary approach as part of the Article 6 on Fisheries Management⁹. The implementation of the Code of Conduct will be facilitated by a series of specific guidelines, one of which will address the precautionary approach to fisheries management (including aspects related to the introduction of new species). The precautionary approach promoted by FAO is being progressively reflected in the fishery sector reality. The applications to inland fisheries and aquaculture have been already mentioned above. In addition, in the last session of the Working Party on Resources Evaluation of the Committee for Eastern Central Atlantic Fisheries (CECAF) it was recommended that, as a precautionary approach, the fishing effort exerted on horse mackerels in Morocco, Mauritania, Senegal and Gambia, should be kept at the level as in the late 1980s. A practical application of the precautionary approach to management of tropical shrimp fisheries has also been proposed (Garcia, 1996) illustrating the possibility to make maximum use of the available scientific information, with its uncertainty, to elaborate precautionary management advice.

More recently, and in direct relation to the process of development of the FAO International Code of Conduct, the Government of Sweden, in close cooperation with FAO, held a Technical Consultation on the Precautionary Approach to Capture Fisheries (Including Species Interaction) in Lysekil, Sweden, 6–13 June 1995 (FAO, 1995). This meeting drafted a set of guidelines (which will support the Code of Conduct) and produced a number of technical background documents dealing in detail with specific technical issues addressed in the guidelines (Fitzpatrick, 1995; Hilborn and Peterman, 1995; Huppert, 1995; Kirkwood and Smith, in press; Rosenberg and Restrepo, 1995). including the present review.

The United Nations process

At its first substantive session, held at New York in July 1992, the UN Conference on Straddling Fish Stocks and Highly Migratory Fish Stocks (hereafter called the Conference) also addressed the issue. It could not reach consensus on the precautionary principle, which many countries equated with a moratorium on fishing and considered too radical for such environmentally soft industries as fisheries. A consensus developed instead on the need to introduce or strengthen the precautionary approach to fishery management. During its Second Session, in July 1993, the Conference considered again the issue. The Chairman negotiating Text (A/CONF.164/13^*) contained only one reference to the precautionary approach, in Article 4: “Use of the precautionary approach shall include all appropriate techniques, including, where necessary, the application of moratoria”. A paper submitted at this meeting by Argentina, Canada, Chile, Iceland and New Zealand (United Nations, 1993) proposed selected precautionary measures on the High Seas, distinguishing between existing and newly discovered fisheries. For existing fisheries, the text suggested inter alia that: (a) TACs and effort limitations shall be established to maintain exploitation rates below the level of MSY and, where appropriate, to allow the stock to rebuild; (b) precautionary management thresholds shall be established at which pre-determined management courses of action should be taken; (c) where stocks decline over time, TACs and effort shall be reduced to arrest the decline and subsidies for fishing operations shall be stopped, and (d) by-catch limitations should be established and stocks of associated or dependent species should be maintained or restored. For newly discovered stocks, the text suggested also that: (a) early large-scale development of fisheries on newly discovered stocks shall be prohibited and limitations shall be applied immediately on effort and on Government assistance, and (b) precautionary Total Allowable Catches (TACs) and quotas shall be established below the MSY level. In addition to these largely technical measures aiming at increasing precaution, the document contained proposals aiming at giving to the coastal States special prerogatives to establish interim management measures: (a) in case of discovery of a new straddling or highly migratory resource and (b) when the coastal State has established that an emergency exists. The heated debate on this latter aspect of the proposal has overshadowed the other aspects of the proposal.

⁹The text of this section (Annex 1) is only provisional and will be revised on the basis of the outcome of the UN Conference on Straddling Fish Stocks and Highly Migratory Fish Stocks

Nonetheless, during its 1993 Session, the Conference requested the Food and Agriculture Organization (FAO) to prepare two information papers: one on the precautionary approach in fisheries management and one on management reference points. During its Third Session, in March 1994, the Conference considered again the issue of precaution, based on the document prepared by FAO and the proposals included in paragraph 5 of the Chairman's Negotiating Text (Annex 2) which referred specifically to the precautionary approach to management. Two working groups were held: on the precautionary approach and on management reference points. The outcome of the heated debate on precaution during the following sessions of the Conference was reflected in a number of modifications of the draft Chairman Negotiating Text which represented a substantial elaboration on the approach (cf. Annex 3 and 4). The UN ad hoc Working Group on Management Reference Points reached consensus on all but one of a set of Technical Guidelines on Biological Reference Points (see Annex 4). The only serious conflictual point, already referred to above, related to the need for coherence in management measures across the area of distribution of the species.

The NGOs process

Non-Governmental Organizations (NGOs), both international and national, environmental or professional have participated actively in the UN process, lobbying for recognition of the need for a precautionary approach to fisheries which would involve, inter alia:

• taking decisions even with inadequate evidence;
• reversing the burden of proof;
• requesting Environmental Impact Assessments;
• avoiding non-reversible impacts;
• adopting management reference points;
• establishing action-triggering thresholds points;
• allowing people's participation;
• promoting transparency;
• establishing sanctuaries;
• taking into account combined stresses on resources;
• reducing by-catch and increasing selectivity;
• conserving also associated and dependant species;
• testing management regimes robustness;
• allowing new fisheries only at very low pilot level;
• establishing dispute settlement mechanisms, and
• promoting inter-generational equity.

NGOs have generally welcomed the FAO efforts towards the operationalization of a precautionary approach to fisheries which recognized the need to: (a) apply it to all fisheries; (b) apply it throughout the stock range, and (c) agree on criteria and actions to be taken before a crisis occurs. Despite complaints of insufficient opportunity for interaction in the Code of Conduct process by some NGOs, it is clear that there is a large coincidence between the NGOs' proposals and the FAO code and guidelines. Some environmental NGOs, however, considered that the FAO approach was too much oriented towards the protection of the fishery sector, making excessive reference to the socio-economic burden associated with it. Some criticized the proposed use of “reversibility” as a criteria for acceptability, considered as a loophole. A fishermen's association, on the contrary, considered that some the FAO proposals were unbalanced, setting an impossible burden for industry. It is clear that more interaction is needed even though there is a basic agreement on what should be done. Expectations of Governments and NGOs may never be identical and differences will also exist between different NGOs. It is therefore probably not reasonable to expect full agreement, by everyone, on all aspects of such a critical issue.

International Council for the Exploration of the Sea (ICES)

Another example of the precautionary approach can be found in the form in which the Advisory Committee on Fisheries Management (ACFM) of the International Council for the Exploration of the Sea (ICES) delivers its advice to its member States. The ACFM states that “for stocks where, at present, it is not possible to carry out any analytical assessment with an acceptable reliability, ACFM shall indicate precautionary total allowable catches (TACs) to reduce the danger of excessive efforts being exerted on these stocks” (Serchuk and Grainger, 1992). The implicit assumption in the ACFM advice is that, in the absence of scientific assessments, uncontrolled fisheries are likely to build up overcapacity and overfish the resources. The preventive action is to establish TACs at conservative levels to limit fishing until better assessments become available. The implication is that such conservative measures would be lifted only if better information, in the form of an acceptable analytical assessment were provided.

In addition to the work on species introductions undertaken with FAO-EIFAC (referred to above under the FAO process), ICES also developed a Code of Practice on the Introduction and Transfer of Marine Organisms (ICES, 1995) dealing more specifically with the introduction of Genetically Modified Organisms (GMOs). It is worth noting in this respect that in considering this Code of Practice, the FAO-SWEDEN Technical Consultation on the Precautionary Approach to Capture Fisheries (FAO, 1995) indicated that “because of the high probability and unpredicted impacts, many species introductions are not precautionary” and that “a strictly precautionary approach would not permit deliberate introductions and would take strong measures to prevent unintentional introductions”.

International Maritime Organization (IMO)

Although not directly related to the fishery sector, the efforts of IMO to reduce the impact of accidental introduction in ballast water and sediment of tankers as well as hull fouling, are worth mentioning. Such accidental introductions are numerous and have resulted in serious damage to the fisheries and aquaculture ecosystem and resources in some cases (Bartley and Minchin, 1995; Mee, 1992; Zaitsev, 1993). The IMO guidelines for Preventing the Introduction of Unwanted Aquatic Organisms and Pathogens from Ship's Ballast Water and Sediments (IMO, 1994) addresses the issue and aim at minimizing the risk of introduction. The issue was also addressed by the FAO-SWEDEN Technical Consultation on the Precautionary Approach to Capture Fisheries (FAO, 1995) which stressed that present practices were largely non-precautionary and that major changes in behaviour, technology and enforcement were required.

The World Conservation Union (IUCN)

The IUCN view on precaution is that “a precautionary approach should underlie all fisheries management, rather than being restricted to special cases” and that “major interventions in the natural environment should not be conducted in the absence of information to assess the potential consequences” (Cooke, 1994). Cooke stressed that it was necessary be not only set and declare the management objectives but also to ensure (through scientific simulations or otherwise) that the management procedures in place result in a high probability to meet these objectives under a wide range of scenarios with respect to stock dynamics and ecological interactions. In order to qualify as “precautionary” a management approach would therefore have “to be sufficiently fully specified to enable its simulation, and to pass at least a minimum checklist of tests”. Cooke, further proposed that authorized levels of catches be inversely related to the amount of data available and that considerations related to protection of fishery habitats, non-target species and biodiversity be included in a precautionary approach. When describing the elements needed to test a management procedure, Cooke lists all the sources of uncertainty regarding the stock, required to predict how the stock might behave (e.g., sampling variability and biasses; uncertainty and long-term fluctuations in stock productivity, dynamics and structure, recruitment, mortality and growth and interactions with other species). Conspicuously lacking from the recommended approach are, however, all the important and often driving sources of uncertainty regarding the fishery sector itself, the fleet and capital dynamics, the alternative employment, the fishermen's behaviour, etc. Without such elements, simulation of management systems in most fisheries would be fairly unreliable.

International Center for the Living Aquatic Resources Management (ICLARM)

The International Center for Living Aquatic Resources Management (ICLARM) has recently developed its position regarding the introduction of species and the need for a precautionary approach (Pullin, 1994) which promotes adherence to the ICES-EIFAC guidelines and acknowledges the potential impact of genetically modified organisms.

Commission for the Conservation of the Antarctic Marine Living Resources (CCAMLR)

While not referring to the precautionary approach explicitly, the CCAMLR Convention includes important principles of ecosystem conservation¹⁰ such as:

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

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

• Prevention of changes or minimization of the risk of changes in the marine ecosystem which are not potentially reversible over two or three decades, taking into account the state of available knowledge of the direct and indirect impacts of harvesting, the effect of introductions of alien species, the effect of associated activities on the marine ecosystem, and of the effects of environmental changes, with the aim if making possible the sustainable conservation of the Antarctic marine living resources.”

¹⁰Conservation taken as explicitly including sustainable use

The last principle is particularly typical of the precautionary approach as it addresses the concepts of risk and reversibility in a broad ecosystem concept (see Kirkwood and Smith, in press) for more details. CCAMLR has also introduced precautionary catch limits for krill fisheries (in 1991 and 1992) and for Electrona carlsbergii (in 1993). It instituted, in 1992, the requirement for advance notification and data requirements prior to the development of a new fishery. Finally, in 1993, in the absence of sufficient data for the establishment of a management regime, it authorized the starting of an experimental fishery for the crab Paralomis spp.

4. UNCERTAINTY, ERROR AND RISK

Uncertainty

In the definition section above, uncertainty has been defined as “the condition of being uncertain. Doubt. Something uncertain. In statistics, the estimated amount or percentage by which an observed or calculated value may differ from the true value” (Houghton Miflin, 1992) or as “the incompleteness of knowledge about the states or processes in nature” (FAO, 1995)

The incompleteness of knowledge derives from: (a) ignorance (i.e., no data at all); (b) inaccuracy (i.e., potential bias in the data), and (c) variance (i.e., statistical confidence limits of the data). More specifically, statistical uncertainty (or variance) is related to stochasticity or error from various sources estimated using statistical methods. In its taxonomy of uncertainty, Wynne (1992) distinguishes between: (a) risk, when the system is basically known and outcomes can be assigned a probabilistic value; (b) uncertainty, when important parameters are known, but not the probability distributions; (c) ignorance: identified lack of knowledge of parameters and relations known to exist and for which are researchable, and (d) indeterminacy: when causal chains and processes are open and thus defy prediction. In decision theory, it is indeed customary to refer to “risk” and “uncertainty” when referring to situations where the outcome of a particular event is unknown, but to use “risk” when the probability of the future event is quantifiable (“knowable”) and “uncertainty” when such probability is unmeasurable (“unknowable”) (Luce and Raifa, 1957; Knight, 1965; Granger and Henrion. 1993). For a discussion on the use of the terms “risk” and “uncertainty” in fisheries, see Shotton, 1993.

In fisheries, the impact of the extracting activity on the resources and the environment needs to be accurately assessed and forecast in order to propose management options reducing to a minimum the possible risk of severe and costly or irreversible crisis¹¹. However, the scientific understanding of the fisheries ecosystems and capacity to predict their future status in accurate quantitative terms is limited by the properties of fishery resources, their “fluid” nature and interconnectedness; the limited knowledge on genetic stock structure and impacts of fishing on resources genetics; the complexity of the interactions between species and gears and fisheries; the poor quality of the available fishery data; the limitation of scientific models and research funds, and the fluctuations of economic parameters. This leads to a degree of uncertainty in the scientific, technical, economic and political information upon which managers and industry leaders base decisions which may not always be wholly appropriate. There are numerous illustrations of this and the most recent relates to the management of the Northern Cod stock in the Northwest Atlantic where, following a collapse of the resources, it was necessary to establish a very expensive emergency welfare programme to support a stunted coastal fishery sector. A polemic has started as to whether research, management, industries, national decision-makers or foreign fleets, were responsible for the mistakes (Finlayson, 1994) and it appears that, as usual, the responsibilities are to be shared and the debate comes too late.

Scientists have repeatedly addressed the issue of uncertainty and the related risk, trying to find ways of identifying and quantifying better the levels of uncertainty in their statements as well as more robust (forgiving) management approaches (Walters and Hilborn, 1978 and 1987; Shepherd,1991; Smith, Hunt and Rivard, 1993). Hilborn (1992) distinguishes between “noise”, “uncertain states of nature” and “surprises”. Noise includes the elements of uncertainty for which historical experience is available, such as year-to-year variations in weather, prices, administration decisions, political setup and directions, etc. and for which probabilities can be usually worked out. Uncertain states of nature refer to elements of uncertainty that have been explicitly identified but for which no experience is available and, therefore, no probabilities can be obtained. These include, for instance, major shifts in ecosystem structure, impact of global change, etc. Surprises refer to elements of the uncertainty that were never considered.

Errors

When decision-makers take the necessary decisions, while both the present situation and the future outcomes are not fully understood, they implicity accept a certain probability to make some mistake and make the assumption that this mistake will either have a negligible cost or would be easily corrected. Errors that might be made may affect: (a) the basic fishery data used for analysis such as on catches, effort, sizes landed, etc. (measurement error); (b) the estimation of populations and parameters derived from such data (estimation error); (c) the understanding of relationships between the different elements of the fishery system and their interaction (process errors); (d) the way these relationships are mathematically represented (model error); (e) decisions that management takes on the basis of such information (decision error), and (f) the way in which management measures are implemented (implementation error). The errors affect both the biological, economic and social component of the fishery system. They may affect, for example, the decision-maker's expectation regarding fishermen's reaction to a proposed measure, as a consequence of errors in the explicit or unformulated behavioural model, used in forecasting such a likely reaction. Management errors can lead to two types of situations:

1. necessary management measures were not taken and, as a result, the resource is damaged. There are short-term costs for the resource and, possibly, for the fishing community if not compensated by government subsidy. The biological impact is usually reversible if a corrective measure is applied, except perhaps in the case of major damage to the habitat. This type of error may also carry the risk of major economic consequences (e.g., in Peru or. more recently, on the Eastern Coast of Canada), and

2. unnecessary management measures were taken and, as a result, fishing activities were curbed. The cost of the error is borne by the fishery. The biological effects of the measure, if any, would usually be positive and reversible soon after the measure is suppressed. The socio-economic impact may or may not be reversible (e.g., where there the error resulted in the loss of the market).

¹¹See a detailed discussion on fisheries impacts in the section on Management Implications

It must, therefore, be accepted that management decisions addressing actual or perceived risks will often be necessarily taken with less than complete and accurate information which may lead to errors. The question is: how to deal with the problem while minimizing the risk of error in the short and long-term? The responses are: (a) improving information to reduce the level of uncertainty, and (b) improving robustness of decision-making to a given level of uncertainty. Improving information and understanding to the point of reducing substantially the risk of error implies data and financial resources requirements which would often be unrealistic, particularly for high seas or highly unstable resources. As a consequence, while research efforts should be pursued, efforts have tobe made to improve decision-making. Hilborn (1992) distinguishes two types of management response to uncertainty. The “blind faith strategies” are based on the best available evidence and applied without any explicit feed-back mechanism for improving them on the basis of performance. These strategies are also called “open-loop strategies” in optimal control theory. On the contrary, “learning strategies” explicitly provide for adaptation and improvement on the basis of more or less active learning gained from experience and surprises. Most management system “learn” but usually do so in a passive or reactive mode, at a very low pace and at the price of costly crises. Active learning would improve performance by accelerating strategy optimization through feed-back loops, and involves “taking management action deliberately designed to be informative in addition to the explicit monitoring and regulation function of management”.

Risk

In the section on definitions, risk has been described as “the possibility of suffering harm or loss. A factor, thing, element, or course, involving uncertain danger, a hazard”. This is the general meaning intended in most environmental conventions. In more technical literature, risk refers to potential negative consequences (or undesirable outcomes) of a decision, quantitatively assessed and often referred to as “expected loss” or “average forecasted loss”. Turner, O'Riordan and Kemo (1991) stress that “risk is not merely an objective phenomenon but a hazard clothed with social meaning and judgement”.

No matter how much effort is made in research and through adaptive learning, acertain level of uncertainty will remain and, therefore, a certain level of risk when making decisions. A fishery management strategy aiming at no risk at all for the resource and the fishing communities would imply either research costs beyond the value of the fishery or no development at all (in the case of an extreme interpretation of the concept of precaution). Few Governments would find either of these two extreme options viable. Cautious management will therefore deal explicitly with risk and aim at a compromise and it should be clear that the higher the uncertainty and/or risk the greater will be the need for caution, particularly in the selection of management reference points (FAO, 1993a). Particular caution may be necessary when resources and people are in a highly vulnerable situation as, for example, in small island countries where the erosion of natural resources may lead to the degradation of the coral reef ecosystem and, beyond a certain threshold, to the breakdown of development opportunities, life support and social order. An important and difficult task for cautious management authorities will be to develop a societal consensus about the nature and levels of the biological and societal impacts (and risks) that might be considered acceptable (tolerable) and to highlight and address the fundamental trade-off implications of the decisions, for different elements of the society and for both the short- and long-terms. Shrader-Frechette (1995) stress that the development of such a consensus would benefit from a science-based comparative risk assessment, to improve the objectivity of possible perceptions of risk and ranking of the various threats to the aquatic system and the fisheries. Such assessment would also help optimize the allocation of human and financial resources available for research, technology development and management. It must be accepted, however, that people are concerned not only with ecological risk, e.g., resource depletion, but also with inequities with regard to risk distribution, lack of concertation on acceptable risks, inadequate insurance or compensation for risk and other non-quantifiable aspects of risk which cannot be easily captured by comparative risk assessment and simple cost-benefit analyses.

Solutions often proposed to the problem of uncertainty tend to be simplistic (e.g. take the “lower bound” of the range) or oversimplistic (discontinue an activity, do not allow it to start), neglecting to compare the cost of this decision to the resulting benefits. Shane and Peterman (in preparation) stress that a precautionary measure “can only be justified if it improves management performance, i.e. if the benefit of reducing overfishing exceeds the cost of reducing harvests”. They suggest whether adjustments to take uncertainty into account are worthwhile and how large they should be.

5. IMPLICATIONS FOR FISHERIES RESEARCH

All expressions of the concept of precaution require that the “lack of full scientific certainty shall be not used as a reason for postponing cost-effective measures to prevent environmental degradation” (Principle 15 of the Rio Declaration). The requirement for precaution may, therefore, have been interpreted as requiring no input from fishery research. Gray (1990), for instance, stated that the “acceptance of the precautionary principle has nothing to do with science” and that it leads to arguments “that do not have the required objectivity and statistical validity”. In practice, however, and as proposed below, the effective implementation of precaution requires substantial support from fishery science, which needs to be adapted to the new requirements.

5.1 The “Best Scientific Evidence Available”

Scientific cooperation to develop a consensus on the state of nature and cause-effect relationships, appropriate models and the potential consequences of fishing has been the basis for cooperation in international fisheries management and the major “raison d'être” of ICES and it should continue to be one of the most neutral contributions to the resolution of conflict between nations and competing user groups. The Christiania Conference, in 1901, held just before the creation of the International Council for the Exploration of the Sea (ICES), endorsed the principle of scientific inquiry as a basis for rational exploitation of the sea. The same principle was also agreed at the International Conference on the Conservation of the Living Resources of the Sea, hosted by FAO (Rome, 1955). The 1982 Convention provided that the best scientific evidence shall be taken into account by the coastal State when designing and adopting management and conservation measures in exclusive economic zones (Article 61). For the high seas, this Convention provides that measures are designed on such scientific evidence (Article 119). More recently, the General Assembly Resolution 44/225 recognized, in its preamble, that “any regulatory measures … should take account of the best scientific evidence available”. The 1982 Convention, however, does not define the evidence required in any quantitative manner.

Regarding the necessary amount of data, Cooke (1994) proposed that there be a relationship between the amount of data available and the level of catches allowed, indicating that a minimum information requirement be established, such as a recent estimate of the low end of the likely available biomass. This might sometimes be difficult to obtain without any fishing at all, although, for many resources, some rough estimate could be obtained through trawl or acoustic surveys. Cooke specifically proposed that “permitted catches be lower when data are sparse than when data are plenty” and stressed that this “attaches a positive effective value to fisheries data and opens the way to data collection programmes financed by the users”.

Regarding the quality of the necessary data, the requirement that the evidence should be the best available implies that even poor evidence can be used in designing conservation measures provided it is recognized as the best available. The 1982 Convention does not provide any guidance on how to decide which is “the best” scientific information. Nor does it indicate how to operate in the absence of a scientific consensus, which it implicity assumes, or when no scientific information is available at all. Although the 1982 Convention does not foresee that an existing fishery could be closed if not enough scientific information is available, it does not impose a great burden to be discharged before the necessary conservation measures can be taken (Burke, 1991). One would assume therefore that, in such a case, the spirit of the Convention is that the missing scientific information should be urgently collected but this does not preclude measures being taken in the meantime. The concept of precaution would ensure that action is not deferred sine die.

Concern has been expressed that the adoption of the precautionary approach could imply that scientific facts to back up management decisions were no longer considered necessary. There is an obvious risk that, by referring to the concept of precaution, scientific objectivity could be less rigorously applied and that international dialogue could be negatively affected. It is hardly debatable, however, that when scientific data are available together with a monitoring and management system, the basic requirement of the 1982 Convention should prevail and decisions should be taken on that basis. It should also be clear that, in order to satisfy the requirement of the 1982 Convention for the best scientific evidence available, the information must be scientific(i.e., obtained and presented in an objective, verifiable and systematic manner)¹² and it does need to be made “available” to all concerned. This, in the context of straddling and highly migratory resources, requires the existence of effective international scientific cooperation and the elimination of non-reporting and misreporting.

In the absence of a scientific consensus, emergency action should, therefore, only be justified when there is the risk of severe and irreversible effects and the concept of precaution may be seen as filling the gaps in the 1982 Convention, preventing the absence of scientific data or consensus from opening a loophole leading to “laissez-faire”management and development strategies with damaging or irreversible consequences. In an international fishery management body, a State willing to invoke the need for a precautionary approach in order to promote exceptionally stringent management measures, would have to convince the other parties that exceptional conditions are met for its application, i.e., that there is indeed a high risk of severe and irreversible damage. Science should, as far as possible, demonstrate the existence and extent of risk through risk analysis. If the available information was considered insufficient to demonstrate objectively the risk, forced application of the concept of precaution could become counter-productive. It is recognized, however,that in such a case, the management authority would have to face “perceived risks”, in the absence of objectively demonstrated ones as is often the case with global societal risks and a consensus will have to be achieved through a largely political process involving as much consultation, participation and transparency in decision-making as possible.

¹²This implies that the “traditional knowledge”, the foundation and accuracy of which is largely unknown, be collected and assessed in order to eventually become part of the “scientific” basis for management

5.2 The Role of Statistical Methods

The 1982 Convention does not give any indications on how to determine which scientific evidence is the “best”. General Assembly Resolution 44/225 required “sound statistical analysis” and this new terminology could be considered an attempt to clarify further the concept of “best evidence”, equating it with “statistically sound evidence”. The advantage of incorporating statistics into the concept is that it offers a way of using well-established mathematical techniques and tests to assess the probability that a certain action has had or may have a certain type of effect. It also forces scientists and decision-makers to recognize and measure explicitly the levels of uncertainty and the risks attached to these decisions. A research programme to monitor a fishery will use statistics to test, for instance, a null hypothesis (Ho) that the ongoing fishing, or planned increase in fishing effort or change in fishing strategy, will not drive (or has an acceptably low probability of driving) the reproductive capacity of the species below some pre-determined safe threshold level. Scientists must still agree on which type of statistical methods to use (parametric, non-parametric, geostatistics) and which test is most appropriate for a particular problem. Fisheries do not usually conform strictly to the requirements for unbiased application of conventional statistical methods and the reliability of many statistical tests might still be a matter for debate. As a consequence, obtaining a consensus on the “best statistical analysis” to use might not always be easy. In this respect, Peterman and M'Gonigle (1992) have stressed the potential contribution of Statistical Power Analysis to the issue. They remind us that “statistical power is the probability that a given experiment or monitoring programme will detect a certain size of effect if it actually exists”. Related to the example given above, it means that the statistical power measures the probability that the fishery monitoring programme will effectively detect the reduction of the reproductive capacity below the safe threshold level. Peterman and M'Gonigle suggest that the lower the statistical power of an experiment, the more precautionary the management response should be. In addition, it is clear that the best statistical methods can only lead to unreliable results if applied to unreliable data. It is, therefore, obvious that rigorous statistical methods should also be applied in data collection systems, particularly for collecting fisheries data.

5.3 The Burden of Proof

The “Proof”

The concept of “burden of proof” is often used in conventions and other texts referring to the precautionary approach. Considering the level of uncertainty which characterizes aquatic systems and socio-economic systems, it should be clear that absolute “proof” stricto sensu is hardly available. The concept, whether of an impact or of the absence of an impact, implies usually a level of certainty that is generally not reachable in fisheries research. In fisheries, the concept of “proof”could be related to the concept of “scientific evidence” established by the 1982 Convention on the Law of the Sea. The “burden of proof”could, therefore, be interpreted as the burden of providing the scientific evidence. It must be noted that just as there is no criteria in the 1982 Convention to define what information is “best”, the references to the “burden of proof” do not provide any guidance as to the “standard of proof” (i.e., the criteria by which to judge whether a “proof” is acceptable). In this respect, the concept of scientific evidence has the advantage to specify that the evidence must be scientific, i.e., obtained and presented in an objective, verifiable and systematic manner.

The Burden

In conventional fishery management, the “burden of proof”, i.e., the responsibility of providing the “best scientific evidence available” required by the 1982 Convention, has fallen traditionally on research and management institutions. It has been necessary for them to demonstrate, with the available data, that the stock could be (or had been) damaged, or that fisheries performance could be improved, before management measures could be imposed. In many instances, this approach has not been effective because fishery research lagged behind development and was not in a position to anticipate changes in techniques and practices. The principle of precautionary action provides a partial solution to this important and recurrent problem in requiring that action be taken even in the absence of “full scientific certainty” about the extent of the risk and the causal relationships. This is often associated with the proposal to “reverse the burden of proof”, i.e., reverse the responsibility to provide the necessary evidence, implying that:

1. human actions should be assumed to be harmful to the resource unless proven otherwise, giving systematically to the resources the benefit of doubt, and

2. the responsibility to prove that human action is harmless or that the impacts are acceptable¹³ lies on those who intend to derive benefits from the ecosystem and not on the management authority.

Proposition (a) may be taken as implying that any fishing technique, which has not been formally authorized, in a given fishery or management area, or for a particular species, is forbidden, a principle enshrined in the FAO International Code of Conduct for Responsible Fisheries. The requirement is related to the notion that an environmental impact assessment should be presented before a new technology or practice is introduced into an ecosystem. It is also related to the concept of prior consent or prior authorization (discussed below Section 6.2). Proposition (b) above, might be more easily implemented in an international agreement, when the party bearing the burden would be a flag State with research capacity. This proposition could, sometimes, be more difficult or impossible to implement at national level when the fishery sector is informal, financially and technically weak or poorly organized as in many developing countries coastal and small-scale fisheries, as well as in overfished fisheries where most of the initiative for corrective action (e.g., fisheries reconversion) starts from governmental initiative.

In most cases a simple Environment Impact Assessment (EIA) based on evidence available locally, or in similar fisheries elsewhere, could be sufficient to produce the evidence required(cf. Section 6.3). In the case of a completely new methodology or fishery (e.g., on a non-traditional species) a major difficulty in the implementation of the concept is that it will be difficult or impossible to forecast, with any degree of accuracy, the impact that the new fishery will have before it has started and some data have been collected. There is, therefore, a real risk that no new fishery could be developed because evidence of the absence of adverse impact cannot be given by those involved in the venture. A reasonable precautionary approach, in such a case, should lead to agreement for a pilot fishery large enough to collect data and build up the scientific evidence required, but small enough to ensure that no irreversible effect is likely¹⁴ (cf. Section 6.4).

¹³For a discussion on “acceptable” impacts, see Section 7.4

An example of application of the concept to international fisheries can be found in the UN General Assembly Resolution 44/225. This resolution recommended a total ban on large-scale driftnet fishing in the absence of scientific consensus on the likely long-term impact, implying that the prohibition of a disputed fishing technique is in order until its acceptability has been demonstrated. It stated that “such a measure will not be imposed in a region or, if implemented, can be lifted, should effective conservation and management measures be taken based upon statistically sound analysis to be jointly made by concerned parties…”. This resolution reversed the conventional course of action. recommending immediate and drastic action (i.e., a total ban of the offending gear) on the basis of international concern assuming that driftnets had an undesirable impact on resources, until shown otherwise. It was agreed that such action could, in principle, be reversed should the joint scientific analysis lead to consensus on the effectiveness of management measures. The UNGA Resolution 44/225 gave no guidance or criteria on how to judge the quality or adequacy of the available evidence or the effectiveness of the management measures. The action was confirmed by General Assembly Resolution 46/215 of 20 December 1991, which called for action against this type of fishing on the basis that “the international community [has] reviewed the best available scientific data and [has] failed to conclude that this practice has no adverse impact … and that … evidence has not demonstrated that the impact can be fully prevented”. Another example of reversal of the burden of proof can be found in Council Regulation 345/92 of the European Economic Community (EEC). which regulated the use and the length of driftnets (limited to 2.5 km) in EEC waters. Article 9(a) granted a derogation until 31 December 1993 to some vessels for the use of longer gear, stating that “The derogation shall expire on the above-mentioned date,unless the Council, acting by a qualified majority on a proposal from the Commission, decides to extend it in the light of scientific evidence showing the absence of any ecological risk linked thereto.”

In addressing the issue of the burden of proof, the Technical Consultation on the Precautionary Approach to Capture Fisheries, held in Lysekil, Sweden, 6–13 June 1995 (FAO 1995), considered that adherence to the guidelines it produced, and particularly to the elements contained in its summary statement (Annex 6), would ensure and appropriate placement of the burden. In addition, the Technical Consultation recognized that the following elements would help clarifying further the issue:

• “all fishing activities have environmental impacts and it is not appropriate to assume that these are negligible until proved otherwise;

• although the precautionary approach to fisheries may require cessation of fishing activities that have potentially serious adverse impacts, it does not imply that no fishing can take place until all potential impacts have been assessed and found to be negligible;

• the precautionary approach to fisheries requires that all fishing activities be subject to prior review and authorization; that a management plan be in place that clearly specifies management objectives and how impacts of fishing are to be assessed, monitored and addressed, and that specified interim management measures should apply to all fishing activities until such time as a management plan is in place, and

• the standard of proof to be used in decisions regarding authorization of fishing activities should be commensurate with the potential risk to the resource, while also taking into account the expected benefits of the activities”.

¹⁴The question is more complicated in the case of introductions of species and GMOs where there is no guarantee that the introduced elements could be safely eradicated once introduced, even on a pilot phase, and there is opposition, in this case to the concept of pilot experiments REF

5.4 Practical Guidelines

In order to support the effective implementation of a precautionary approach to fisheries management and development, fishery research needs to be adapted to the new requirements and should, in particular:

1. ensure that the “lack of full scientific certainty shall be not used as a reason for postponing cost-effective measures to prevent environmental degradation” (principle 15 of the Rio Declaration);

2. take into account the best scientific evidence available when designing and adopting management and conservation measures, in accordance with the provisions of the 1982 Convention;

3. require a minimum level of information to be made available for any fishery to start or continue;

4. make all necessary efforts to collect the required scientific information. For new fisheries, data collection should start with the fishery, including data on genetic and stock structures. For existing fisheries, data collection should start as soon as possible and any increase in effort should be preceded by a research or assessment programme;

5. ensure and require that information provided as a basis for management be “scientific” (i.e., obtained and presented in an objective, verifiable and systematic manner) and “available” to all concerned;

6. develop the effective international collaboration required to collect and jointly analyse the scientific information, particularly inthe case of trans-boundary, highly migratory or high seas resources;

7. take measures aiming at eliminating or reducing non-reporting and misreporting, inter alia, by ensuring that the fishery sector cooperates in data collection and is fully informed of the results and uncertainty in the assessment;

8. relate the allowance in terms of TACs, catch quotas, number of licences, etc. to the amount and quality of the available data, ensuring that permitted catches be lower when data are sparse rather than when data are plenty;

9. generalize the use of standard statistical procedure to judge the quality of the scientific evidence available and ensure that such information and the analysis therein is statistically sound;

10. improve statistical methodologies for assessing the biological and economic parameters, testing their sensitivity to uncertainties in the data used and systematically estimating bias and precision in the derived parameters. The sensitivity of models to uncertainties in their parameters and functional structure should also be tested;

11. assess the statistical power of the tests and methodologies used for comparing the relative “soundness” of the information available. The lower the statistical power of the assessment, the more precautionary the management measures;

12. develop standards of proof and agreed protocols for Environmental Impact Assessment, pilot projects and experimental management projects;

13. promote multidisciplinary research, including: (a) social and environmental sciences, and (b) research on management institutions and decision-making processes, because the availability of biological evidence alone has not prevented overfishing;

14. expand the range of fishery models (e.g. bio-economic, multi-species, ecosystem and behavioural models), taking into account: (a) environmental effects; (b) species and technological interactions, and (c) fishing communities' social behaviour;

15. systematically analyse various possible management options using the whole range of available models, showing: (a) the likely direction and magnitude of the biological, social and economic consequences, and (b) the related levels of uncertainty and the potential costs of the proposed action (risk assessment), and no action (status quo scenarios);

16. systematically analyse and highlight the most pessimistic scenarios¹⁵, in situations of doubt and high risk of irreversible damage to the resource;

17. develop scientific guidelines and rules for multi-species and ecosystem management as a basis for agreement on acceptable degrees of disturbance;

18. agree on quantitative reference points and thresholds as well as on methods to establish them¹⁶;

19. systematically quantify the risk associated with scientific advice at the various reference levels selected;

20. improve understanding of environmental impact, raising the awareness of fishermen to the possible impact on fisheries potential resulting from fisheries as well as from environmental degradation caused by other industries, and

21. improve technological research on fishing gear and practices and their environmental impact.

¹⁵For instance, models which assume strong dependence of recruitment on adult stock size and predict rapid collapse when effort develops beyond a critical level (such as the Gulland-Schaefer production model or the Ricker stock-recruitment model), should be used rather than models assuming no relation between stock and recruitment and high resilience of stocks to high fishing rates (such as the Fox production model or the Beverton and Holt yield-per-recruit and stock-recruitment models)

¹⁶For instance, if it is agreed that it is safe to exploit a resource at two thirds of its MSY, it will be necessary to agree on the reference data set and on the conventional model on which to base the calculations because the true value of 2/3 MSY, and of its corresponding level of effort, will never be exactly known and may vary according to the model used

6. IMPLICATIONS FOR TECHNOLOGY DEVELOPMENT AND TRANSFER

Fishing affects targeted stocks and associated species, reducing their abundance and spawning potential, changing size structure and species dominance or composition and modifying the trophic chain. These effects are “normal” in the sense that they result from the need to exploit fish, and must be addressed and kept at acceptable levels by management (see Section 7.4). Fishing also has side effects on the flora and fauna living in the exploited environment (birds, turtles, marine mammals, benthic communities, coral reefs, seagrass beds) as well as on the bottom itself (trawls and dredges). In addition, “ghost fishing” by lost or discarded driftnets or pots has been suspected and, in some instances, demonstrated. It is not by chance that the very first discussions, in FAO, on the concepts of responsible fisheries, focused on responsible “fishing”, i.e., on responsible fishing gear and technology, before broadening the concept to cover also management, research, fish processing and trade and aquaculture.

An example of international concern is given by the reaction to the rapid expansion of the large-scale pelagic driftnet fishing (see Section 5.3). The problem has been apparently “solved” by a moratorium on all driftnets of more than 2.5 km in length, through heated debate and political wrestling, but Miles (1992) indicated that the application of the same flawed process and criteria to EEZ fisheries would lead to closing down of many of them¹⁷. Another example is the concern expressed regarding impacts on cetaceans off Ireland and Denmark (Schoon, 1994) by bottom gillnets of up to 7 miles long, used in coastal waters, for the last 15 years to catch bottom fish such as turbot, plaice and cod.

The following sections, which draw from the work of Boutet (1995), will address various ways in which the problem could be addressed in the context of a precautionary approach to fisheries, i.e., through the adoption of responsible fishery technology and practices, the establishment of technology lists, the adoption of Prior Informed Consent and Prior Consultation Procedures, the requirement for Environmental Impact Assessment and the implementation of pilot or experimental development projects.

6.1 Classification of Responsible Fishery Technology

In international environmental law, the precautionary principle is often associated with the requirement to use the “best available technology”, an obvious parallel to “best scientific evidence available”. This wording has sometimes been interpreted as requiring the technology which has the smallest environmental impact, regardless of the short-term socio-economic costs. This interpretation has, however, been questioned on the basis that such technology might not always be affordable by all countries and, in particular, by developing countries (GESAMP, 1986). General Assembly Resolution 44/228 of 22 December 1989 on UNCED referred instead to “environmentally sound technology”, stressing the need for socio-economic constraints to be taken into account. The wording does not pretend to limit the choice to a single “best” or soundest technology and does not preclude, therefore, the use of many “sound” technologies together, depending on the socio-economic context of their introduction. The Cancun Declaration (Mexico, 1992) provides that “States should promote the development and use of selective fishing gear and practices that minimize waste of catch of target species and minimize by-catch of non-target species”, focusing on only one of the challenges of responsible fishing.

¹⁷As a matter of fact, arguments similar to those used to request the closure of the large-scale pelagic driftnet fisheries were invoked to force the closure of the small-scale bottom gillnet fishery in California, showing both the potential and the danger of media-driven campaigns against fishing techniques

The development of typologies and classifications is usually the basis of a process of normalization or standardization of technology in view of its regulation. The basis of a classification in fisheries could be horizontal or vertical. A vertical classification would involve classifying gears according to their priorities with the aim to regulate their use. An horizontal classification would classify ecosystems and species assemblages, or parts of them, as a basis for the regulation of their use. In practice, both classifications would be required in order to develop flexible regulations taking into account the diversity of gears and ecological situations (and even socio-economic situations). The use of lists to classify chemical substances, techniques, species¹⁸, weapons, etc. is fairly frequent. In environmental law, technologies are often catalogued on separate lists, the “colour” of which reflects the perceived degree of environmental friendliness. For instance:

“Black” or “Red” lists would identify technologies for which the likelihood of producing unacceptable impacts in most or all of their application.

“Grey” and “Orange” lists would identify technologies susceptible to produce potentially acceptable impacts in most of their applications but which should be used under some conditions and require a specific impact assessment before being introduced.

“White” or “Green” lists would identify those technologies believed to be harmless or producing only acceptable levels of impact and which could be introduced without a particular precautionary procedure.

The task is not easy. One problem is in deciding whether one would catalogue gear, aid to navigation and detection (which increase fishing power) or fishing practice, or both. Another problem is to decide on the objective criteria for the classification. If responsible fisheries is the objective, gear should be classified according to related criteria (referring for instance to selectivity and by-catch rate; impact on bottom, navigation and environment in general; relative energy consumption; biodegradability; difficulty to control and monitor, etc.). For fishing gear, the classification of a technology will depend, inter alia, on the type of habitat. Heavy trawls may be considered“green” on deep muddy grounds but “red” in shallow estuaries and coastal zones or coral reefs. Artificial reefs might be on a grey or orange list because their impact on coastal habitat is long-lasting and, if made of derelict material, they may contaminate the environment.

This list approach has been indirectly applied to fisheries by reference to the Convention on the Conservation of European Wildlife and Natural Habitats (Bern, 1979). The Convention gives, in its Annex IV, a list of non-selective gear to be banned, which includes all nets. Although it had been designed for migratory birds, the list has been referred to, in Italy, in connection with the banning of large-scale pelagic driftnet fishery. The importance of nets in fisheries and their contribution to the livelihood of small-scale fishermen and indigenous people illustrates the need for careful consideration before referring to lists contained in non-fishery agreements and before elaborating specific lists for fishery technology.

¹⁸CITES, has recorded species in lists, according to their status, and specific measures correspond to each list

Considering that, in fisheries, the concept of responsible fishing is well defined and that a Code of Conduct for Responsible Fishing has been prepared and will be adopted, it may be of value to refer to the requirement for “Responsible Fishery Technology” (including capture and post-capture technology) as defined in the Code and its different guidelines. Responsible technology will have to be used in all areas of fisheries, including capture, land-based or sea-based processing and distribution. As a consequence, although some general guidelines can be given, based on known characteristics of types of resources and technology, the most responsible mix of technologies to be used in a particular fishery will have to be agreed on a case-by-case basis with explicit reference to the agreed management reference points and acceptable levels of impact agreed for that fishery. The implication is that technology lists could not be for general application and would have to be established locally, at regional and national level.

One must recognize, however, that lists of prohibited gears and practices exists in most national legislations and that these are frequently ignored. Examples are: fishing with dynamite or poison, fishing with scuba-diving equipment, use of obstructive shaffers on trawls cod-ends, use of driftnets, of small-meshed beach-seines, etc. The efficiency of technology classifications and list of authorized gears is therefore strongly dependant on the capacity of monitoring and enforcement.

Care would also have to be taken to ensure that the use of gear lists does not lead to freezing the evolution of technology and that mechanisms exist (including the use of pilot projects) to allow this evolution while keeping the overall fishing mortality under control. Fitzpatrick (1995) also stresses that, in many instances, the technology necessary for fishermen safety, also improves the fishermen's ability to locate and catch fish and, therefore, contributes to overfishing. Such technology, often required by international conventions on safety on board of fishing vessels cannot however, in most instances, be removed from the vessel. The implication is that fleet size may have to be reduced when fishermen safety is improved, in order to stabilize fishing mortalities.

Moreover, a “better” technology might be theoretically available on the market but in effect not accessible to some countries because of its cost or its sophistication and, in many instances, the generalization of the use of responsible technology will require an improvement in international cooperation in technology transfer, as underscored in Agenda 21¹⁹.

6.2 Prior Informed Consent (PIC) and Prior Consultation Procedures (PCPs)

For dangerous polluting industries, reference has often been made to Prior Informed Consent (PIC) and Prior Consultation Procedures (PCPs). The practical significance of the procedures involved is that, before introducing a dangerous technology or any new technology in a controlled or sensitive area, the proponent must produce a substantial amount of information about the technology to be introduced and its potential impact and, eventually, obtain the consent of the State or the managing authorities. If the introduction is agreed, a number of specific measures are usually foreseen such as limiting the scale of the initial project, special monitoring and reporting requirements, etc.

These practices are rare in fisheries. An example can be found in the ICES/EIFAC Code of Practice to Reduce the Risk of Adverse Effects Arising from Introduction and Transfers of Marine Species including the Release of Genetically Modified Organisms (Turner, 1988) which has been adopted by the International Council for the Exploration of the Sea (ICES) and the European Inland Fishery and Advisory Commission (EIFAC) of FAO. The ICES/EIFAC Code foresees that “Member countries contemplating any new introduction should be requested to present to the Council, at an early stage, information on the species, stage in the life cycle, area of origin, proposed plan of introduction and objectives, with such information on its habitat, epifauna, associated organisms, potential competitors with species in the new environment, genetic implications, etc., as is available. The Council should then consider the possible outcome of the introduction, and offer advice on the acceptability of the choice.”

¹⁹The successful efforts made by the Inter-American Tropical Tuna Commission in the Eastern Central Pacific area to train crews of the region in effectively avoiding by-catches of dolphins through the use of appropriate technology, is a good example of what can be achieved in this respect

The European Directive 90–220 on dissemination of genetically modified organisms intends to frame the development of biotechnologies in Europe and address the “genetic risk” potentially represented by these technologies, which are of great potential interest also for fisheries (EEC, 1990). Hermitte and Noiville (1993) stress the precautionary character of the Directive, which applies the precautionary principle, not to a single product (chemical substance), or to a specific problem (ozone hole), but to a whole new mode of production, even before any incident has been registered. The Directive recognizes that a new production mode carries with it significant social (societal) changes and potential risks and, contrary to what has happened in industrial development since the 18th century, attempts to foresee and limit the negative impacts of this new technology. It reverses the traditional industrial culture and freedom to undertake, produce and sell as long as a danger has not been proven.

In exclusive economic zone fisheries, where effective effort controls have been established, there is often a requirement to obtain prior consent from the management authority before a new vessel is ordered or even before the banks are approached for a loan for this purpose. A similar approach might be used for some particularly efficient and potentially dangerous technologies and/or for particularly vulnerable resources or fragile ecosystems when severe, irreversible effects are possible. In a regional or international context, Prior Informed Consent of the competent regional management organization or arrangement would be required before introducing a new methodology. The procedure would be better accepted if the new technology was patented, limiting the risk that the benefits to the “discoverer” could be jeopardized in the process. In such an international or regional mechanism, a State willing to introduce a new technique would be requested to present a report, comparable to an Environmental Impact Assessment (see section on EIA below). Such an assessment would address potential effects on the target species, on associated species which might be targets for other fisheries in the area or food items for such target species and on the environment.

It has been mentioned that an overly stringent application of the precautionary principle might be contrary to the willingness and need to ensure technological progress. Hermitte and Noiville (1993), however, indicate that the prior authorization process, the resulting direct involvement of industry in promotion of data collection and research, and the transparency resulting from the public information and participation would, on the contrary, contribute to dissipate the fears towards technology and, indeed, limit irrational reactions to innovative technologies. One major benefit from a prior authorization process, beyond the limitations of risk, would be in the mandatory delivery, by industry, its scientists and experts, and at industry's expense, of information on ecosystem functioning and technological impacts and of the resulting “memory” that Hermitte and Noiville call “scientific jurisprudence”. These authors state that the acceptance of the procedures by scientists and industry would be a sign of good faith given to a more and more suspicious, sceptical and unforgiving society and that these procedures may in fact be the only way to avoid irrational bans on research and development avenues and the development of “wild” experiments.

The administrative burden imposed by prior authorization procedures could be overwhelming and, at least in fisheries, there would be obvious advantages if the procedure could remain exceptional. The scope of application (and unnecessary burden) of the measure could be reduced using the concepts of “familiarity” and “previously acquired experience” (Hermitte and Noiville, 1993) or referring to “evidentiary presumptions” (Bodansky, 1991) to take into account available knowledge obtained elsewhere in similar or sufficiently comparable conditions, to reduce the amount of uncertainty and presumption of risk. In order to avoid repeating the impact assessment of similar technologies on similar species and ecosystems, it would be useful to develop a general typology of fishery technologies, gears and practices and their potential impact, leading to a general impact-oriented classification of gear/species/ecosystems interactions, to be used as a guide, by management authorities, at regional or national level, to develop local gear and technology classifications based on local characteristics of the resources and the environment²⁰(see also Section 6.1). The special monitoring and reporting procedures could then be limited to new technology/species/ecosystem combinations and to existing technologies recognized as unacceptable in the long term and for which phasing out might have been decided (and for which interim reports could be requested during the phasing out period).

In the case of high seas areas not covered by any specific international agreement, there would be no competent authority to which the request for prior consent could be made. In addition, there would also be no monitoring or enforcement system in place, making it impossible to detect the introduction of harmful techniques and to measure impact. This is a case where the legal responsibilities of the flag States would need to be clearly determined, especially if the flag State registers all vessels authorized to fish in the high seas as provided for in the 1993 Agreement on the Promotion of Compliance with Conservation and Management Measures by Fishing Vessels in the High Seas.

6.3 Environmental Impact Assessment (EIA)

Impact assessment is a major instrument of environmental law, which conditions the beginning of an activity or the deployment of a technology to an assessment of the consequences on the environment. Generally, an EIA provides not only an assessment of the impacts but also proposals aiming at mitigating the impact if necessary. As it would not be practical to condition all fishing activities to EIA it might be necessary to define the conditions under which an EIA might be necessary. This could be done: (a) through preliminary studies, on a case-by-case basis, and (b) through an overall identification and cataloguing of the technology/resource combination requiring such approach (see above).

The EIA seems to have been rarely used in fisheries (except possibly in aquaculture and for species introductions). If generally adopted, the EIA procedure would be part of the legal procedure leading to the granting of a fishing right or license for a particular fishing activity by an authority with the legal competence required to authorize or deny such a right. This authority would define the requirements and specifications of the EIA. An EIA procedure would require the establishment of a system to control the conditions of the assessment, its relevance and objectivity. This implies that:

²⁰This comparative approach is not really new in fisheries, but the process of fisheries law development, in developing countries, to which FAO contributes actively, involves already a lot of transfer of experience from area to area. The approach could however be formalized and more systematically applied

• the proponent would be allowed to appeal if the procedure imposed is not in line with the established specifications, or if the decision of the authority does not appear in line with the conclusions of the EIA;

• the authority, which would decide on the acceptability or otherwise of a new technology or practice, would have to be able to oversee the whole EIA process to guarantee to all users the quality and reliability of the assessment;

• the procedure should be transparent to all users who receive information on request and on the EIA process. It might be necessary to organize a debate on the issue to have all views. It would be essential to ensure that the authority keeps the necessary prerogative to ultimately decide;

• the other users (and in particular the users of a different technology on the same resource) should have the possibility to appeal on a decision if it appears to be in contradiction with the conclusions of the EIA, and

• as a last resort, recourse to tribunals (in EEZs), or to dispute settlement mechanisms (in international fisheries), should always be possible if one of the parties in the EIA process believes that its interests are being unduly affected.

There should be some relation between the cost of the EIA and the cost of the potential negative consequences of the proposed development and its potential benefits. There should also be some relation between the cost of the foreseen investment and the cost of the EIA. In some instances, participation by the authority or State in the EIA might be worthwhile and equitable, particularly when the technology being considered has general potential application. State participation in the EIA would certainly be necessary for coastal and small-scale fisheries, particularly in developing countries (see Section 5.3 on the burden of proof).

6.4 Pilot Projects

Despite their relatively smaller size, fishery pilot projects can be considered a s“full-scale” experimentations, only limited in duration and geographical extension. They could be a useful way to implement a precautionary approach to fishery development provided that specific rules are adopted for their conduct, data collection, and analysis. They have the advantage of being less theoretical than EIAs, and therefore more convincing, while limiting the probability of inadvertently damaging the resource, and allowing a more realistic approach to socio-economic impacts than otherwise possible. Allowing for a phased approach to application of technology at a larger scale, they represent a practical tool for implementation of a “stepwise decision making” and “progressive deconfinement” of a new technology, advisable to situations of high uncertainty (Hourcade, 1994). Pilot projects have been extensively used in the past, including in FAO fishery development programmes, to demonstrate the technical and economic feasibility of a development or management measures. An experimental fishery has been developed for instance on Paralomis spinosissima crab fishery in the Antarctic (CCAMLR area) (Watters, 1993) and the concept is one with which industry is generally familiar. A basic assumption behind the concept of pilot projects is that the large-scale implementation of the technology is a simple extrapolation of the pilot scale. This may not always be the case and a significant involvement of basic and applied sciences is necessary for improving the protocol and specification of traditional pilot projects allowing them to become also useful and reliable elements of a precautionary fishery development policy. Another implicit assumption is that all traces of the experiment can be eliminated if the pilot-scale project indicates that the tested approach or technology results in unacceptable consequences. This may not always be true and explains the opposition of some scientists to the concept, particularly in cases where the consequences detected in the pilot project are not reversible (as may be the case with introduction of GMOs). The implication is that only part of the cost of a pilot project could be considered as additional charge required for precaution. Most of it could, in many cases, be considered as normal pre-investment expenses.

The management authority should have enough latitude to impose, to a proponent of a new technology or new fishery, the type of experimentation considered most appropriate. A contractual agreement between the authority and the proponent would improve the probability that the rights of the “discoverer” of a technology or a stock are respected.

The pilot project goes beyond the EIA in the sense that real development will occur, even though at small scale. In some cases, the authority itself could be (and often has been, in the past) the promoter of the initiative. In some cases, both an EIA and a pilot project might be required and executed sequentially when the EIA is not totally negative but some aspects may not be addressed without experimentation.

6.5 Practical Guidelines

A precautionary approach to fisheries should ensure the use of responsible fishery technology in all sub-sectors, including capture, land-based or sea-based processing and distribution and ensure that:

1. technology, formally recognized as “responsible”, is compatible with long-term resource conservation, minimized by-catch of endangered species and discards, as well as other non-acceptable impact;

2. the mix of responsible technologies (and practices), to be used in a particular fishery, is agreed on a case-by-case basis with explicit reference to the management reference points and acceptable levels of impact agreed for that fishery. This mix should be compatible both with local conditions for sustainability and socio-economic conditions of the operators;

3. recommended technologies are easily available on the market and affordable for developing countries and that their transfer is promoted through international cooperation;

4. criteria for the selection or determination of responsible technology include local biological and environmental conditions and socio-economic constraints;

5. selection or determination of responsible technology is based on an objective assessment of the actual or likely impacts and of the risks involved, for the resources, associated species and, in the long term, for the fishing community, taking into account the type of resources, ecosystem characteristics, and habitat;

6. technological requirements are defined with a view to maintaining (or reducing) the accidental effects of capture and post-capture fishery activities within pre-defined acceptable (tolerable) levels, allowing general application by all countries or parties involved;

7. States and management organizations and mechanisms undertake to list the fishery technology used or potentially usable, the “colour” of which would reflect the perceived degree of environmental friendliness;

8. before introducing a new technology in a controlled or sensitive area, on a low-resilience or particularly vulnerable species, the proponent is asked to produce a sufficient amount of information about the technology to be introduced and its potential impact and that the prior consent of the other users is required when appropriate;

9. if the introduction of a new technology is agreed, a number of specific measures should be foreseen such as limiting the scale of the initial project, special monitoring and reporting requirements, etc.;

10. when adopting PIC or PCPs, States or regional management, organization or arrangements should ensure that the potential rights (interests) of the inventor of the resource or of the technology can be protected;

11. request for the introduction of new techniques be supported by documentation amounting to an EIA identifying potential effects on the target species, and on associated species, which might be targets for other fisheries in the area or food items for such target species;

12. PIC and PCPs procedures should remain exceptional in order to reduce the administrative burden imposed to fishermen, and

13. special monitoring and reporting procedures should also be used for activities recognized as unacceptable in the long term and for which phasing out has been decided. Interim reports could be requested during the phasing out period.