A. INTRODUCTION

A first major review of the state of world fishery resources was produced by FAO more than two decades ago (J.A. Gulland (ed.), 1971. "The Fish Resources of the Ocean". Fishing News Books, England, 255p), and since then the FAO Fisheries Department has been producing periodical updates almost every two years. While these reviews and their updates are primarily intended as background documents for the regular sessions of the FAO Committee on Fisheries (COFI), they have proven to be most useful as background documents for other international meetings dealing with the conservation and management of marine living resources, as well as to individual fishery scientists and managers looking for a brief but comprehensive review of the state of world fishery resources.

The present document should be largely regarded as an update of the 1993 Review of the state of world marine fishery resources, issued as FAO Fisheries Technical Paper No. 335, and of its previous update, issued in 1995 as FAO Fisheries Circular No. 884. However, in addition to a new layout, this review provides information on fish production over an extended time series (1950-1994).

The analysis of this extended time series, now readily available in electronic form, has led to some interesting findings and general conceptual postulates regarding the general trends and potential of fisheries at global and regional scales. Some preliminary results of the analysis of these general trends are provided in the following two sections. A more detailed description of the state of exploitation and management by major regions will follow, to end with a section on special topics of general interest.

Each regional review is split into three main sections: the "Introduction", a "Profile of Catches" and a section on "Resource Status and Management". The "Introduction" describes the key geographical features and key attributes of the physical environment found within the specific FAO statistical area. In the introductory section there is also a brief summing up of the main points of current interest or concern within the FAO statistical area, whether they arise from long-term trends or patterns or recent changes in the fisheries. The "Profile of Catches" draws on the long-time series of catches on data recently collected into a single database by FAO to give a sense of the historical scale, development and relative importance of the various types of fisheries found within each FAO statistical area.

However, the main focus of each regional review is on recent developments within specific fisheries as described in the section on "Resource Status and Management". The structure of this section within each regional review varies to reflect the most appropriate basis upon which to separate the resources into meaningful components. In certain FAO statistical areas, separation into Exclusive Economic Zones is used, while in others a distinction is made according to resource type (e.g. demersal, pelagic). In other areas authors have followed the approach used by organization(s) charged with assessing or managing marine resources in that specific area. Wherever possible, the results of stock assessments are used to make quantitative comments on resource status directly. However, in many fisheries, particularly those which have developed recently, such information is not available. In such cases qualitative information may be presented. Additionally, in instances where catches are believed to be indicative of stock status, quantitative catches data may be discussed. The description and analyses of state of exploitation given in these sections make use of the best information currently available to FAO.

For each FAO statistical area a table containing catch data for each year from 1988 to 1994 and ten-year averages from 1950 to 1989 is given in Section D "Marine Resources Tables". Also contained in these tables are annotations on the state of exploitation of each resource (see the "Notes for all Tables" in section D for definitions). A certain degree of care must be taken in interpreting these annotations because they are given on a species-by-species basis, rather than for individual stocks. It is typically the case that within each FAO Statistical Area the catches of a given species will come from two or more distinct stocks. It is often the case that the state of exploitation of such separate stocks is different. In such instances the "state of exploitation" indicators in the table should be checked against the main narrative within the appropriate "Resource Status and Management" section for further details. Given this distinction between species and individual stocks, it can be seen that the "state of exploitation" notes have limited statistical significance from a fisheries management point of view and are intended as "rule of thumb" indicators only. It should also be noted that, while FAO believes that the principal marine species which are, or have been, the subject of exploitation are represented in the tables, the majority of species which contribute to many marine fisheries in lesser quantities are not included.

1. A PERSPECTIVE ON FISHERIES TRENDS OFFERED BY ESTIMATES OF PRODUCTION PER SHELF AREA

The growing use of Geographical Information Systems (GIS) technology and databases has allowed estimates of shelf area to be made for FAO statistical areas. Fishery production figures, otherwise expressed in terms of landings, can therefore be made more easily comparable by expressing them, as for agricultural and forestry resources, per surface area available for their production. When landings of shelf-dependent resources (bottom fish and invertebrates) and small- and medium-sized pelagic fish (living resources associated with, but not restricted to, shelf waters) are each expressed per surface area of continental shelf within the 200m depth by individual FAO statistical areas (Tables A1.1 and A1.2), certain surprising regularities emerge. At least half of the marine statistical areas realized peak production figures at some time previous to the last five-year period of 1990-94, (especially for areas in the Northern hemisphere), and these peak values show marked similarities within given ranges of latitudes. It is also possible in this way to compare peak production values between areas subject to different climatic and oceanographic conditions.

As far as the shelf resources are concerned, there is a marked similarity in peak production per shelf area between FAO statistical areas in the northern hemisphere, where former peak production levels clustering around 2.7 t/km2 were realized, and comparable arcto-boreal areas in the Southern Hemisphere, where peak levels were lower at some 2.15 t/km2 . The exception to this pattern was the Southeast Atlantic, whose peak levels were much higher during intensive exploitation by distant water fleets in the 1960s and 1970s.

Statistical areas which lie in the tropics also show a similar clustering of production figures, with peak production levels for shelf resources that are much lower, at some 0.74 t/km². Similar figures were obtained for the Mediterranean (0.70t/km²). Obviously, there are 'hot spots' within each area where local production levels are much higher, such as estuaries and coral reefs, but the average productivity over the whole shelf appears more to reflect natural restraints due to nutrient supply, especially once fishing effort has exceeded optimal levels.

The role of nutrient supply seems to be most evident with respect to pelagic resources, where peak production figures vary from 0.5 t/km² for areas with stratified, nutrient-poor water masses to as high as 25 t/km² in FAO Statistical Area 87, where the most important upwelling-driven pelagic fishery in the world occurs. Peak pelagic productivity is clearly associated with upwellings in the tropics and sub-tropics but, interestingly enough, upwelling systems are not especially conducive to high production of shelf invertebrates and bottom fish, probably because of environmental instability and the low oxygen conditions that strong upwellings occasionally lead to. In cold-temperate seas, high pelagic production is associated with tidal mixing of nutrient-poor surface and nutrient-rich bottom waters, especially on the extensive continental shelves of the Northern hemisphere.

Obviously, there are tropical areas (such as the Western Indian Ocean) where significant increases in production of tunas and mesopelagic fish may be realizable given strong upwellings, but for most other tropical areas, other than those with locally high production areas associated with coral reefs and estuaries, further increases in production are constrained by nutrient supply. Such nutrient supply is generally poor in stratified tropical waters, and fishery productivity usually drops off rapidly with depth. Here, despite a relatively short history of intensive fishing, there appears to be limited future potential for further increases in production per shelf area, and if there such a potential, this seems likely to come from further exploitation of small pelagic resources and tunas.

2. DEVELOPMENT TRENDS AND POTENTIAL

Figure A2.1
During the period considered by the first estimate of world potential production based on analysis of historical landings made in FAO by J. Gulland¹ in 1971, landings were increasing at about 6% per year (Figure A2.1), and Gulland estimated that the potential for traditionally exploited marine species was about 100 million tonnes per year. This estimate of fishery potential was consistent with estimates made earlier by several other authors (see Moiseev², p. 203), based on different analyses. In fact, the growth rate for marine production observed by FAO soon fell, although some growth was maintained. Despite fisheries developing on non-traditional species, the marine fishery production has so far only reached about 90 million tonnes (in 1994) with capture fisheries accounting for 84 million tonnes.

Summarised here are two aspects of a recent study³ of fishery development trends and fisheries potential based on the analysis of FAO landing statistics for the period 1950-1994 detailed by species and major fishing area, about double the extension of the series of data which was available to Gulland.

Trends in fishery development

Figure A2.2
The process of development of a fishery as described by changes in landings with time has been described by many authors (e.g. Caddy and Gulland⁴, Caddy⁵). This process is schematically represented in Figure A2.2 as comprising four phases: (I) undeveloped, (II) developing, (III) mature, and (IV) senescent. The Figure shows the theoretical change in yield and relative rate of increase of yield during the development process. The relative rate of increase, which varies significantly as the maximum long-term yield is approached, reached and "overshot", is of particular interest and has been used here to provide a rough assessment of the state of marine resources, globally and by ocean. This rate is nil for a stable non-developing fishery (phase I), but increases rapidly (phases I-II) as the fishery starts to develop. It then decreases during the phase of steady growth of the fishery (phase II) and drops to zero again when the fishery reaches its maximum production (phase III). In reality, this representation is of course obscured by natural fluctuations as well as by noise in the data. However, the trend in decline of the relative rate of increase during phases II and III was used here to estimate when full potential has been reached and what the corresponding potential multispecies yield could be.

It is important to note, however, that aggregate landings from various stocks which are the subject of a fishery-complex may continue to increase despite local overfishing situations as long as the process of increase through expansion to new areas and resource elements overshadows the process of decrease through overfishing. An important implication is that the highest landings observed in a "mature fishery" represent a sort of multi-species composite average long-term yield (ALTY) which is different from the sum of the theoretical maximum sustainable yields (MSYs) for the various resource elements, whether these elements represent different species in a given area or even different areas with their multispecies resources components in a given region or ocean.

Figure A2.3
The top 200 species-area combinations for marine fish, referred to here as "resources", were selected for analysis on the basis of average landings over the whole time period. These 200 major resources, which account for 77% of world marine fish production, were grouped by cluster analysis, according to the shape of the landing trends, irrespective of the scale of the landings. Twelve resource groups were defined, and for each of these an average landings trend curve was described. These curves can be considered as comprising parts of the overall fishery development model described above and reflecting some of the various phases identified (from I to IV, from undeveloped to senescent). Each curve was divided into segments corresponding to different phases of the generalised model (Figure A2.2) and the number of resources in each phase of fishery development calculated for each year. The overall pattern of change in the phase composition is shown in Figure A2.3.

This figure strikingly illustrates the process of intensification of fisheries since 1950 and the increase in the proportion of world resources which are subject to declines in productivity ("senescent", phase IV). It also underlines the fact that the ever-growing total tonnage of world fishery production gives a misleading vision of the state of world fishery resources and a false sense of security (a comment already made by FAO⁶). Unfortunately, a similar comment should probably be made for changes in total aggregated landings at national level, so often used as a justification for further development.

The results shown for 1994 indicate that about 35% of the 200 major fishery resources are senescent (i.e. showing declining yields), about 25% are mature (i.e. plateauing at a high exploitation level), 40% are still "developing", and 0% remain at low exploitation (undeveloped) level. This indicates that around 60% of the major world fish resources are either mature or senescent and, given that few countries have established effective control of fishing capacity, these resources are in urgent need of management action to halt the increase in fishing capacity or to rehabilitate damaged resources. A strikingly similar conclusion was reached by FAO⁶, which concluded that 44% of the stocks for which formal assessments were available were intensively to fully exploited, 16% were overfished, 6% depleted, and 3% slowly recovering, concluding therefore that 69% of the known stocks were in need of urgent management. That same study, using a global production model, with estimates of the world capacity, concluded that the demersal high-value species were overfished and that a reduction of at least 30% of fishing effort was required to rebuild the resources.

Potential of marine fisheries

In a review of the state of world fisheries undertaken in 1994, ⁶ showed that the annual relative rate of increase of world reported landings had significantly decreased since 1950 and was approaching zero, indicating that the maximum production from the world's conventional marine resources under current exploitation regimes was being approached and that the mean catches of the last few years were probably very close to that maximum. The same analysis has been repeated below globally and by oceans.

Based on the analysis of total marine landings, the predicted maximum production for world marine fisheries with the present overall fishing regime (generally characterised by small sizes at first capture and significant discards) corresponds to about 82 million tonnes, a value close to the average landings of 1990-94 of about 83 million tonnes. This crude global estimate of the world potential of conventional resources under the present regime of exploitation indicates that the world would be at full potential about now. This is a composite, aggregate result, which hides the increasing occurrence of overfishing on a multitude of stocks in many different areas, as evidenced by global statistics over the last half century. In order to reduce this effect, and illustrate differences in development rates, timing, as well as potential, the analysis has been undertaken separately for each ocean.

Figure A2.4
The time series of landings for the Atlantic, Pacific and Indian Oceans, as well as for the Mediterranean Sea (including the Black Sea) have been processed in the same way as the world marine total landings, and the corresponding lines showing the declines in relative rate of increase are shown in Figure A2.4. Only the main conclusions are reported here and details can be found elsewhere3.

  Estimates of world marine production potential were obtained from the two analyses as follows:

a) - Estimate based on world marine total landings (Potential = 82 million tonnes);

b) - Summing individual estimates made for each Ocean (Potential = 100 million tonnes).

The aggregated estimate of 82 million tonnes refers to the present fishing regime. The difference between the aggregate potential (in case a above) and the sum of the different oceans' potentials (in case b above), which amounts to 18 million tonnes, would come partly from optimisation of the production in the North Atlantic and increasing landings from the Indian Ocean. However, this latter possibility results from a rather unreliable extrapolation and should not be used as a basis for planning further development in this ocean.

In order to gain some insight as to how fishery potential might be distributed within each ocean, the same analysis was undertaken at a lower level of aggregation by considering total landing for each of the 16 FAO major fishing areas for statistical purposes. The results are summarised in Table A2.1. Taking all the results at face value, the analysis indicates an overall potential of 125 million tonnes, assuming that each area can be separately optimised. However, the statistical fits vary and in some cases are much poorer than for the analysis by ocean. It is therefore strongly recommended that this section be read in conjunction with the regional reviews in the following section. A subjective classification of the reliability of each of the estimates of potential is given in Table A2.1.

The most reliable estimates (marked as ** in Table A2.1) yielded by the above analysis would indicate that an increase of 10 million tonnes from recent landings of about 83 million tonnes could well be possible (i.e. leading to a possible total potential world marine production of 93 million tonnes), including 4 million tonnes coming from improved management of overfished stocks in each of the Atlantic and Pacific Oceans and 2 million tonnes from fisheries development in the Indian Ocean.

If less reliable estimates (marked * in Table A2.1) are taken into account, an additional increase of 17 million tonnes (i.e. leading to a possible total potential world marine production of 110 million tonnes) might be possible, comprising 2 and 1 million tonnes from management of overfished stocks in the Atlantic and Pacific Oceans respectively and 14 million tonnes from development in the Pacific Ocean. However, this further possible increase is more problematic.

Further, taking into account the least reliable estimates would lead to an additional increase of 15 million tonnes from fisheries development in the Indian Ocean, leading to a total potential world marine production of 125 million tonnes. This perspective remains, however, highly uncertain. These hypothetical potential increases, particularly the last two, corresponding to total production of 110 and 125 million tonnes, have to be considered with prudence because some critical regional assessments (i.e. Indian Ocean, Mediterranean and Southeast Pacific) are not reliable and require further analyses to confirm or reject the preliminary results given above.

The analysis predicted that full potential for the Indian Ocean had not been reached and will not be so for decades (Figure A2.4), and the estimate of potential should be considered very unreliable until further substantiated. Generally speaking, the Indian Ocean has an important biological potential but its coastal resources are under severe stress in many areas and require effective management, even though potential for expansion may exist further offshore. The scarcity of stock assessment work in the area makes it difficult to confirm or reject the result of this preliminary analysis. It must also be pointed out that, although the possibility of expanding Indian Ocean fisheries offshore and into deeper grounds has been repeatedly considered by some coastal countries in that area, particularly India, Indonesia, Malaysia and Thailand, no convincing evidence of high additional levels of resources has been provided yet (except perhaps for tropical tunas and tuna-like species).

It is interesting to note that the estimates of the possible gains to be obtained through management or development depend on the level of aggregation in the analysis. The more aggregated the data used for the analyses, the less is the estimated potential for further increase. This is an expected result, as only a disaggregation of the time series by region and stock can reveal the various situations of under-development or overfishing that exist simultaneously in all areas. This analysis of regional resources takes the disaggregation only some way towards the stock level of detail. It is also worth noting that the potential estimated at the ocean level of aggregation (100 million tonnes) lies between the potentials estimated from the analysis by major fishing area utilising only those estimates considered reasonably reliable (93 million tonnes) and those considered reasonably reliable and less reliable (110 million tonnes).

Implications for management and development

For the resources which are presently below their historical peak levels of production, it might be possible to return to these levels by reducing fishing effort and, in most cases, simultaneously improving yield-per-recruit. This can be achieved by increasing significantly the age at first capture, prohibiting the exploitation of juveniles, increasing mesh sizes, and closing temporarily or permanently areas of concentrations of young fish. Examples in Cyprus and the Philippines⁷ have shown that increases of sustainable production of 100% can be obtained in the tropics within 18 months. More recent experiences with the protection of juveniles in Morocco (on cephalopods, through a closed season-area) and Norway (on cod, through ad-hoc area closures) have also produced improvements in catch rates which tend to show that short-term benefits can also be expected in more temperate areas. Effective management can undoubtedly also lead to long-term increases in yield. This has been demonstrated for the Northeast Arctic cod stock in the Norwegian and Barents Seas which, in contrast to most other Atlantic cod stocks, has shown a recovery in spawning stock biomass from a depleted condition to a level not seen since the 1950s following a major reduction in fishing mortality in the late 1980s⁸.

An important problem and opportunity is in the potential improvement from the reduction of unwanted bycatch. It has been estimated that between 18 and 39 million tonnes of fish are discarded every year⁹ (average = 27 million tonnes), comprising of species of low commercial value but also of a large proportion of juveniles. These are part of the catches, if not part of the landings and are not included in the data used in this study, but nevertheless need to be taken into account for management. If added to the recent landings, they result in a world marine catch of about 101 to 122 million tonnes (average 110 million tonnes). The benefits resulting from a reduction of unwanted bycatch through increased survival of juvenile fish can be very significant.

Increases in production would come from further fisheries expansion on those resources which are apparently still increasing their contribution to world landings (about 40% of major fish resources are classified as still "developing" in this study and FAO in 1994⁶ estimated 32%). It must be recognised that even though for these resources landings are still growing at a steady rate, it is not possible to have a reliable estimate of the potential. It would be a mistake, however, not to recognise that some potential exists.

An important question is whether, at the global level at which the analysis has been conducted, improvements in yield from both demersal predators and pelagic prey can be expected. The pelagic group contains a number of significant predators, among which are the large pelagic tunas and tuna-like species. The eggs and larvae of many demersal species, and the adults of some small demersal fish, are prey for the small pelagic species. Similarly, small pelagics are often major dietary components for larger demersal fish. The implications of these interactions are not easy to foresee and it is therefore impossible to establish how much of the present balance in the abundance (and potential) of pelagics and demersals results from the relative overfishing of the demersals and the resultant reduced pressure on pelagics. Neither is it possible to determine to what extent the rehabilitation of the overfished demersals will affect the survival and potential of the pelagics. The issue of resource rehabilitation at large regional scale has never been tackled and remains, with the issue of the medium-term variations of the small pelagics, one of the key issues of the management of fisheries for the 21st century.

In conclusion, while it must be recognised that the statistical significance of most of the analysis described here is insufficient, the elements of information available indicate that an increase in fisheries production of at least 10 million tonnes is possible plus further increases in landings of an unknown magnitude obtained from fisheries development as well as from mariculture. FAO in 1995¹⁰ indicated that an additional 20 million tonnes more of landings might be obtainable. The results of the present study provide a firmer basis for believing that such an increase can be realised if: (a) degraded resources are rehabilitated, (2) under-developed resources are exploited further, avoiding, however, their overfishing and avoiding the overfishing, of those resources which have already reached the highest level of sustainable exploitation they can with stand, and (3) discarding and wastage are reduced.