Interactions between tuna fisheries: A global review with specific examples from the Atlantic Ocean

Alain Fonteneau
ORSTOM
213 rue Lafayette
75480 Paris Cedex 10 France

Pilar Pallares Soubrier
Instituto Español de Oceanografia
Corazon de Maria 8
Madrid, Spain

ABSTRACT

This paper has two goals. The first is to develop some general considerations about the issue of interactions between tuna fisheries; the second is to review interaction problems in the Atlantic Ocean. The general consideration include a discussion of the types of tuna migrations, since migration is a key component in the interaction phenomenon. The potential importance of a “cryptic biomass”, i.e., a fraction of the tuna stock with low catchability, is stressed. It appears that for various tuna stocks this cryptic component may be of major importance; it may temporarily overestimate the exploitation rate of tuna stocks and subsequently overestimate the interaction between fisheries. A biological classification of the various tunas species, from tropical (e.g., skipjack) to temperate (e.g., bluefin) is developed and discussed. Each tuna species has developed a specific type of migratory pattern between the spawning and feeding zones. The migratory patterns appear to be of key importance in producing various types of interactions.

The various types of interactions, in the short term and in localised areas, and in the longer term (yield per recruit) are also reviewed and discussed in this paper. Various interaction problems still pending are discussed, such as the specific problem of interactions in fishing areas located in the extreme range of the stock distribution, or the interaction problems in relation with the possible homing behaviour of spawning tunas. This homing behaviour may be an important biological characteristic for various tuna species such as yellowfin, bluefin and possibly bigeye. The types of interactions observed or suspected between the Atlantic tuna fisheries are reviewed for all tuna species. Although the problem of interactions was never explicitly discussed by the scientists of the International Commission for the Conservation of Atlantic tunas (ICCAT), it appears clearly that large potential interactions probably exist in the Atlantic Ocean for various tuna species. However, due to the heterogeneity and complexity of the interaction problems, it appears that most of the present interaction results are still preliminary and hypothetical. These uncertainties and difficulties found in the Atlantic Ocean are probably similar to the ones encountered in most other oceans.

1. INTRODUCTION

The tuna fisheries in the Atlantic Ocean have shown an increase in activity since the early 1960s. This is well shown by the global increasing trend in the catches observed for the major tuna species (Figure 1); the increased catches are the consequence of increased effort. This increasing trend was observed to various degrees for most species, except albacore and western Atlantic bluefin (Figure 2). The analyses conducted every year by the Standing Committee for Research and Statistics (SCRS), the ICCAT scientific committee, have also indicated that most of the Atlantic tuna stocks have already reached a stage of full exploitation and some are considered overfished. In such conditions, it is quite clear that the problem of interactions between the various tuna fisheries should be of increasing importance. The interactions studied in this paper will be the potential effects of one fishery on the potential catches and catch rates of other fisheries, catching tuna in the same area or in different fishing zones. The problem of interactions between fisheries is of a general nature, as described by Ricker (1958), but the tuna fisheries and tuna stocks are probably a peculiar case, primarily because of tuna migrations, the wide areas covered by most stocks, and the diversity between the various gears catching various ages of tuna.

Figure 1. Catches of tuna fisheries in the Atlantic Ocean.

As a high exploitation rate is always a key factor involved in the interactions between fisheries (Shomura et al., 1994), the increasing catches may also increase the interactions between the Atlantic tuna fisheries. However, the SCRS of ICCAT has paid very little attention to the problem of interactions between fisheries. This apparent lack of concern by ICCAT may be that, in the Atlantic Ocean, very few developing coastal countries are catching tunas with an artisanal fishery. As a consequence, fishery managers rarely pressure ICCAT to pursue interaction problems.

Figure 2. Catches of the major tunas and swordfish in the Atlantic Ocean.

Interaction problems are indirectly tackled by the SCRS at a global and wide scale, using multi-gear yield per recruit models (Ricker model). In this type of model, the fishing mortalities of each gear are first calculated (following VPA and based on the catch by age for each gear), and then the potential for yield per recruit competition between gears are calculated. This type of analysis has been conducted since the early 1970s on yellowfin tuna (Lenarz, 1971), and more recently applied for other tuna species. However, this global yield per recruit analysis (conducted under a hypothesis of a homogeneous unit stock) is unrealistic for tuna and tuna-like species. For instance, it is now obvious that in the Atlantic Ocean (and probably in all other oceans) the gear yield per recruit analyses conducted on yellowfin caught by the purse seine and longline fisheries do not realistically reflect competition between the two gears. These analyses always indicate that the longline fishery should provide a better yield per recruit than the purse seine fishery, when the real fishing performances of the longline gear is always quite limited. Fishery data from all oceans show that longline gear is inefficient to catch large amounts of yellowfin, at least compared to purse seines.

Very few stock assessment models or, more generally, stock assessment works targeting potential interactions between fisheries have been presented for the Atlantic Ocean and discussed by the SCRS. However, this type of modeling could be well addressed in the Atlantic, at least for some species. The basic data, which are always of key importance in those analyses, are the detailed catch by size and effort data by gear statistics (by detailed time and area strata). These data are available in the ICCAT database for most tuna species. However, the interaction analysis often requires significant tagging of tunas in order to provide realistic evidence of interactions. Preferably, the results should include good recoveries as a very low recovery rate is, most of the time, impossible to interpret. Tagging programs have been conducted in the Atlantic, but only on a limited scale since the general lack of funding in ICCAT during recent years has limited tagging programs. Most of the Atlantic tagging experiments were conducted many years ago, under different fishing conditions and often with low exploitation rates. The results of those studies may be difficult or impossible to extrapolate to the present fisheries and stocks. Furthermore, the various limitations of tagging and recovery data imply that complementary methods such as genetic studies or microchemistry of bones should be conducted to evaluate better the true stock structure of tunas.

The goal of this paper is to develop some general thinking on the interaction problem between tuna fisheries at a world-wide level, and then to examine the present interaction problems, or potential problems, for the major tuna species caught in the Atlantic.

2. GENERAL THINKING ON TUNA INTERACTION PROBLEMS

2.1 General Thinking on Tuna Migrations

Tunas are a world-wide distributed family. Various genera and species from this family have a common biological feature of being “highly migratory”. This biological feature has been introduced in the Caracas Law of the Sea meetings, giving to this family a special status in the basic Law of the Sea text (Article 64). The text notes that fishes classified as highly migratory should be managed outside the national EEZs because of their highly migratory patterns. Fish migrations in general and tuna migrations in particular have been extensively described and analysed by many authors, including for the Atlantic Ocean papers by Harden Jones (1981), Cayré et al. (1988) and Cayré (1990).

Based upon the biological cause and type of movement, most tuna migrations can be classified into two categories as "feeding migrations" and "spawning migrations". The extent and types of those migrations vary between species. However, the term “migration” has been often used in an inappropriate manner to describe any movement of tunas; a good scientific definition should be much more restrictive than the general concept of migrations used by lawyers in Caracas.

The types of tuna movements can be described as “advective” or “diffusive”. The “migration” sensu stricto corresponding to an "advection type" movement, is a massive and oriented movement towards a biological target, such as a spawning ground or a feeding ground. The homing behaviour often observed for migratory species that show a strict fidelity to spawn in their area of birth belongs to this category. This definition could, for instance, well be applied to southern bluefin tuna migrating, year after year, from their spawning zone south of Java to specific feeding zones of the southern oceans (Figure 3).

Figure 3. An example of migrations between feeding and spawning areas : the southern bluefin case. Average southern bluefin catches by the longline fisheries during the historical period (average 1956-78). The areas with high average catches correspond to show the sub-equatorial spawning areas (south of Java) and the feeding areas (southern latitudes).

A "diffusion” type movement is a movement of tunas within their ecologically favourable zone. As this favourable zone may often show a seasonal movement in various fishing zones of the Atlantic (for instance in the Senegal area, Figure 4), this seasonality of the environment often produces a massive seasonal diffusion of the tunas, inside the favourable zones. This diffusion usually follows the apparent geographical displacement of their ecologically favourable environment (for instance the 22°C isotherm for tropical tunas, or the 18°C isotherm for albacore).

The time and space heterogeneity of the environment plays an important role in most of those tuna movements. Both the global movement patterns of tunas and their variability, are under the control of the environmental conditions, because the time and areas strata favourable for spawning and for feeding are limited and variable in time and space. However, it appears that the physiological adaptations of the various tunas species to the environment, especially the thermoregulatory devices developed by most tuna species, are quite different (Sharp, 1978; Brill, 1994). Consequently, each tuna species has developed a specific physiological and ecological strategy. This diversity of adaptive physiology has allowed the tuna family to colonise most parts of the epipelagic zones of the world oceans. These evolutionary and ecological components may play an important role in the exploitation of those stocks, for stock assessments by scientists, understanding the interactions between and among fisheries, and for the management of the tuna resources.

2.2 Tuna Stocks and Cryptic Biomass

2.2.1 The concept of cryptic biomass

A major biological characteristic of most tuna stocks is that their distribution and movements cover wide geographical pelagic zones. Consequently, most tunas have in common, most of the time, a large potential “cryptic biomass”, i.e., a fraction of the stock which is not available to the existing fisheries. This non-available biomass may simply be due to the large oceanic distribution of most of the tuna species; the distribution is often wider in volume (horizontally and vertically) than the strata exploited by the fisheries. This “cryptic biomass” with low exploitation rate shows, most of the time, an incomplete mixing with the biomass in the exploited area. The same problem may be found either geographically (Laloe, 1989; Die et al., 1990), or vertically (depth stratification) when the species exploited by surface gear also inhabits the deep layers of the oceans as is the case with most large tunas. Consequently, the cryptic biomass is either stable or decreasing moderately, in comparison with the heavily exploited fraction of the stock located in the main fishing zone.

This cryptic component may disappear when the fisheries extend their range of operation vertically and geographically. However, a situation where 100% of the total biomass is available to the fisheries may be quite rare, as most tunas cover large oceanic zones and are scattered between the surface and deep waters. Not accounting for this cryptic fraction of the stocks may have been responsible for a major component in error by scientists in their stock assessment analyses; multiple examples of this error can be found for various species in various oceans. The consequences of this error was that, when a tuna fishery starts to develop in coastal areas, the scientist tends to conclude (a) from production model studies the stock was reaching its MSY and the fishery was at its optimal effort and (b) from tuned VPA studies the level of fishing mortality was already high in relation to the low recruitment level.

These erroneous conclusions can easily be explained by the fact that, within the heavily exploited coastal area, (a much smaller area than occupied by the total stock), the potential productivity is smaller than the productivity of the stock would provide, thus the local abundance decreases very quickly as does the local CPUE under a local high fishing pressure (Boggs, 1994; Fonteneau, 1991).

Figure 4. Example of diffusion seasonally driven by the environment: seasonality of tuna fisheries and local sea surface temperatures in the area of Cape Vert (Senegal, period 1980-91). Yellowfin, skipjack and bigeye tunas being caught primarily in warm waters (in the area with sea surface temperature >24°C), are forced to migrate seasonally in relation to the marked seasonality of the environmental conditions.

(a) Average sea surface temperature in the Cap Vert area (source: merchant ship data).

(b) Average monthly baitboat catches in the Cap Vert area, by 1° of latitude.

2.2.2 Working hypothesis: “A tuna stock is not a swimming pool”

In a swimming pool, a person located in any place around the pool, equipped with a small bucket will be able to remove all the water from the swimming pool. The volume of water removed from any place in the pool will correspondingly decrease the level of water everywhere in the pool (Figure 5 a). If a tuna stock was similar to the water mass in a swimming pool, any localised fishery could be able to produce an over-exploitation of the total stock; this was never observed for tuna fisheries.

Figure 5. Why a tuna stock is different from a swimming pool?

The decrease of water level in a pool when water is taken at any site of the pool.

Changes of relative local biomass when a fishery exploits a given geographical fraction of the tuna stock (in this case, the left fraction of the stock).

In a tuna stock, covering large oceanic areas, most fishery data show that, when a fishery operates only in a small fishing area, its catch decreases the overall stock biomass much less than the biomass within the fishing zone itself (Figure 5 b), corresponding to a local depletion of the biomass. As an example of this fundamental rule (still hypothetical, but probably a realistic assumption for most tuna stocks), the local biomass of Atlantic swordfish will be more depleted inside the US EEZ, which is intensively fished, than in the unexploited central Atlantic, where significant biomasses of swordfish are also found. Based on the Japanese longline catch statistics, a significant biomass of swordfish exists in the central Atlantic but is little exploited for economical reasons. This difference is due to the incomplete mixing of the various fractions of the stock between those two areas, at least in the short and medium time periods. It is a common feature for most tuna stocks that only the areas with higher densities of tunas, and the areas more easily reached by the fishing vessels, are fished significantly. Unknown and potentially large fractions of the stock thus often remained unfished. The relative importance of this unexploited fraction of the stock may be a function of various factors, including:

(a) Economics: Areas of low densities are not exploited for economic reasons. As an example, most albacore stocks will have a greater fraction of their stock unexploited than bluefin tuna, primarily due to lower fishing effort relating to the lower value of albacore. Most albacore stocks can only be exploited by longliners in the strata with high densities. On the contrary, fishermen can exploit bluefin tunas at low fish densities and still make a profit because of the high value of bluefin in the sashimi market.

(b) Gear limitations: When the gear available to catch a species is not very efficient (this is often the case with the gear catching albacore), various fractions of the stock may remain unavailable. This component may be changed by improvements in fishing technology (e.g., size of vessels, electronics, etc.).

(c) Concentration: Ecological factors, such as the dispersion and concentration patterns of the species and stocks are important factors in determining the unexploited fraction of the biomass. A scattered biomass is always more difficult to catch than a concentrated one. It is well known that all tunas stocks have relatively low biomass per unit area due to the large zones they inhabit. Thus, the average density is always quite low and the resource can be fished economically only when fish are concentrated in specific areas, e.g., in feeding or spawning zones.

(d) Mixing: When the mixing rate is high and permanent between the various geographical components of the stock, the stock may become extensively exploited and the cryptic fraction of the stock may be insignificant. This case (similar to the swimming pool concept) seems to be quite rare, or even never observed, for tuna stocks. In most cases, the mixing seems to be quite incomplete (horizontally and vertically) and consequently the effects of any localised fishing is more apparent at a local scale (for instance, within an EEZ) than at the global scale of the stock. This hypothetical rule was implicitly accepted by Hilborn and Sibert (1986) when they discussed the EEZ management of skipjack in the Pacific Ocean. Various existing migration models which can simulate the migratory pattern of tunas exploited by various fisheries (e.g., the models of Kleiber and Baker, 1987, or Deriso et al., 1991) could be used to evaluate this fundamental problem.

Where a moderately exploited stock is considered overexploited, most calculations of interactions between fisheries would conclude that there is a significant potential for interactions because of the high exploitation rate, which is always a key component producing the interactions. In fact, the true fishing mortality may still be quite low and the total population very large, so the potential for interaction may be lower than calculated.

When the fishery develops horizontally (extended fishing zones) and vertically (for instance, by developing the use of deep longlines), the stock assessment analysis provides much revised estimates of the MSY, optimal effort and recruitment level. These three parameters are generally dramatically higher than the original estimates. The changes of the yellowfin tuna fisheries have been well analysed and modelled by Die et al. (1990) and Laloe (1989) for the eastern Pacific and eastern Atlantic, respectively. Under this development of the fisheries leading to a fully exploited stock, the revised estimates of MSY (and other parameters such as F) are improved estimates of the stock parameters, and estimates may then be converging toward the true parameters of the stock.

The opposite bias and error can also be observed when a fishery, after a period of overfishing, shows a decreasing trend of its fishing effort which is often associated with a decrease in the range of the fishing operations. This case is quite rare presently, but may be observed for species such as albacore. As noted earlier albacore are difficult to catch with modern fishing gears or are not sought due to market reasons (too low value). In such a case, the reduction of the fishing zone may tend to produce an “artificial over-exploitation”, with decreased MSY and decreased recruitment levels. Such a situation is due to the reduction in size of the fishing area; the decrease is not due to a stock collapse.

The changes in the estimate of MSY and recruitment levels should not be interpreted as biological changes in the exploited stocks. For instance the increased recruitment calculated from VPA studies in the early 10970s for the eastern Pacific and eastern Atlantic yellowfin fisheries were probably linked with an expansion of the fishing effort targeting those resources. A careful review of the scientific literature in the field of tuna stock assessment may show other examples of this type of apparent changes in resource dynamics.

Consequently it should be very important to estimate the volume, age composition and distribution area of the cryptic biomass in order to make accurate stock assessment estimates and to identify the importance of interaction. The movement patterns of tunas between the unexploited and exploited fractions of stock should also be estimated, as the role of this cryptic biomass will be very important when mixing is relatively low; on the other hand the cryptic biomass may be neglected when the mixing rate between areas is intense and permanent.

2.3 Biological Classification of the Major Tunas

2.3.1 General patterns

Tunas belong to a family which shows well how a systematic group can develop an evolutionary diversity starting from a common root. This common root of tunas originated in the warm tropical waters of the Miocene and Pliocene eras, more than 10 million years ago. Most tuna species have retained some aspects of their geological origin such as their tropical locations for spawning areas and nurseries.

Additionally, most tunas have developed thermoregulation capabilities (Sharp, 1978; Brill, 1994); this thermoregulation is due to various sophisticated anatomical features which are specific to tunas. The range in efficiency of thermoregulation is, however, variable and dependent on size, as thermoregulation facilities are proportional to size. While all tuna species are able to colonise colder waters when they become larger, the ability to swim deeper and/or at higher latitudes is dependent on size.

Based on size and thermoregulatory capabilities, tuna can be classified into three groups (Figure 6). These are:

Group A includes tropical species, such as skipjack tuna and Auxis spp. The species included attain small maximum size and are predominantly distributed in warm waters (>22° C) during their entire life.

Group B includes tropical-temperate tunas such as yellowfin and bigeye tunas. These tunas show extensive migrations between their spawning areas in equatorial waters and their feeding grounds in the southern and northern latitude. This biological pattern is especially true for adult bigeye which are distributed deeper and at more temperate latitudes than yellowfin.

Group C includes temperate tunas such as bluefin tunas. These tunas show very extensive migrations between their spawning areas located predominantly (but not always) in warm waters, and their feeding grounds which are located in temperate or cold waters.

Figure 7 summarises the thermoregulation pattern for the various tuna species. The tropical tuna species of Group A are predominantly of small size (usually less than 5 kg), have high natural mortalities (M near 1.0), short expectation of life (less than 5 years), and consist of large numbers of individuals recruited yearly in their populations (in the order of billions of fishes).

On the other extreme, the temperate tunas of the Group C (their prototype being the bluefin tunas) show opposite characteristics. These include large adult sizes (reaching often 500 kg), low natural mortality (at about M=0.2), long expectation of life but late spawning, usually after 5 years, and small numbers of fishes recruited yearly (hundred of thousands yearly).

The intermediate Group B (yellowfin, albacore and bigeye) show intermediate characteristics for all those biological parameters. The three group show a striking similarity in the relative distributions of their feeding and spawning areas (Figure 6).

Group A (tropical tunas), is distributed in large tropical areas suitable for them to spawn and to feed. Spawning often occurs all year round in most warm waters areas; areas where the tunas encounter conditions suitable for sexual maturation. Maturation of these tunas is usually very fast and spawning is generally opportunistic. As those areas are often poor feeding grounds, the species in Group A often move towards the colder and richer subtropical areas, adjacent to the spawning areas, where they can temporarily feed better and grow faster than in warm equatorial waters. This group of tunas show predominantly diffusion type movements.

Figure 6. A simplified ecological classification of tunas: spawning and feeding zones.

(a) Group A. Diffusion-type movements, latitudinal and longitudinal, in a non-stationary, environmentally-favourable zone. A large, equatorial spawning area that is slightly smaller than the area of distribution.

(b) Group B: Large but specific equatorial spawning areas, and large migrations of adults in subtropical and temperate waters (often deep).

(c) Group C. Dislocation between warm subtropical spawning areas and remote temperate feeding areas. Precise migratory behaviour and strict homing, between small spawning sites and temperate or subpolar feeding areas.

Figure 7. Major biological characteristics of various tunas species.

(a) Schematic overview of the changes of optimal water temperature for various tunas, versus their age (BFT=bluefin, BET=bigeye, YFT=yellowfin, SKJ=skipjack).

(b) Schematic overview of the approximate minimal sea water temperature for various tuna species, versus the maximum weight of each species and the average levels of recruitment (BFT=bluefin, BET=bigeye, YFT=yellowfin, SKJ=skipjack).

Group B (yellowfin, albacore and bigeye) shows more geographical segregation (in time and space) between their original spawning grounds and the northern and southern (or western/eastern) feeding grounds. Spawning takes place usually in specific equatorial areas during limited periods, when mature fish are concentrated for spawning. Feeding grounds for the adults may be located at great distances from the spawning areas; those tunas may spend a large proportion of their life cycle outside the spawning zone. Massive and large scale migrations often takes place between the spawning and feeding grounds covering hundreds or thousands of miles; this extensive movement is well shown for instance for the Atlantic yellowfin (Figure 8). Similar large scale movement may occur with yellowfin in the Indian Ocean.

Figure 8. The transatlantic migrations of yellowfin tunas between feeding zones (west Atlantic) and spawning zone (Gulf of Guinea) shown by transatlantic recoveries.

Group C (bluefin) represents the most extreme case of “geographical dislocation” between the spawning and feeding zones. Spawning usually takes place in warm waters, but always within very specific time and area strata. The bluefin will spend most of its life outside the spawning area; they are the only tuna able to feed in the remote and rich cold waters of the northern and southern temperate areas. This biological behaviour definitely implies a strong homing behaviour. The southern bluefin tuna has adults scattered in the Indian, Pacific and Atlantic Oceans and spawns south of Java. This species must have an efficient homing behaviour in order to return in due time to their remote birthplace from far distant places. The same situation will apply to the other bluefin tunas (Atlantic and Pacific). All bluefin tunas stocks probably offer various examples of strong homing behaviour which provide good examples of the "obstinate nature" in migrations as developed from an evolutionary point of view by Cury (1994), it is clear that the bluefin tunas should have a consistent and strict spawning fidelity in order to be able to reproduce efficiently in a precise short time and area window. Group C, and to a lesser degree Group B, provide the best examples of the “true” tuna migrations which involves massive and directed movements.

In summary, the small tropical tunas in Group A have developed a biological strategy which involves hundred of millions of small individuals exploring, more or less randomly, the entire warm inter-tropical area, feeding and spawning on an opportunistic basis, and following the seasonal movements of inter-tropical warm waters.

Large temperate tunas in Group C have developed an opposite biological strategy where small numbers of very large tunas are able to migrate over great distances and to explore and exploit efficiently the cold northern and southern temperate feeding areas. The species use their larger size for better thermoregulation and more efficient migrations. This strategy will necessarily imply from an evolutionary point of view an efficient capacity to return to their individual tropical birth strata (“homing”). This is a necessary key to the survival of those species.

The species in Group B show intermediate life strategies in terms of number of individuals, maximum sizes attained and migrations between spawning and feeding zones. The homing component in group B may also be an important component. However, the advective and diffusive components in their movements may vary between oceans (the spawning migrations are not well known for the yellowfin in the eastern Pacific) and more research should be conducted on this important topic.

The strong homing behaviour for some tuna species may have been neglected by tuna scientists. This homing behaviour may have definite implications in stock assessment and management of various heavily exploited tuna stocks, such as most of the bluefin stocks and possibly some of the other species.

It may consequently be of major importance to better estimate the genetic heterogeneity within each stock (or management unit), know the fidelity of specific groups of tunas to given migratory patterns and consequently the potential interactions between fisheries located in various areas and to determine the fidelity of each sub-population to its birth strata (homing).

3. GLOBAL CLASSIFICATION OF THE TUNA FISHERIES INTERACTIONS (based on experience in the Atlantic Ocean).

3.1 Atlantic Tuna Fisheries: Data, Fisheries and Environment

Very few tuna interaction problems can be considered as being ocean-specific. However, various features in the Atlantic may be quite unique and help facilitate interaction studies. The features include:

(a) the small size of the Atlantic Ocean, which is more or less a closed geographical unit and covers only 90 millions square km;

(b) tuna fisheries have shown a strong development during recent years, starting from more or less virgin stocks, to fully exploited fisheries (see Figure 1);

(c) an oceanographic environment which has been well studied by various oceanographers; the Atlantic is characterised by a strong seasonality in most areas (temperate, tropical and equatorial) and a low interannual variability; the Pacific, on the other hand, shows low seasonality and marked year to year variability (El Niño).

3.2 Long Term Interactions

The problem of recruitment overfishing, or how overfishing can affect the recruitment levels, is an important aspect of interaction. It may explain how overfishing the spawning stock may reduce the recruitment levels and the potential catches of another fishery directed to catching juveniles. However, the typical stock recruitment model, such as the models proposed by Beverton and Holt (1957) or Ricker (1958) do not apply to tropical tunas, at least for most tropical species. For these species it seems that the variability in recruitment is only moderate from year to year. No trends are apparent even at high levels of exploitation of the adult stock as experienced in the Atlantic (ICCAT, 1994a). In fact, a decreasing recruitment trend due to overfishing (recruitment overfishing) has never been observed for any tropical tuna, even for highly exploited stocks.

Consequently the present paradigm for Atlantic tropical tunas can be that, at moderate or relatively high level of fishing efforts, there will not be any effect or interaction of the adult fisheries on the juvenile ones. This specificity of tropical tunas may be related to the extremely high fecundity of these individuals (female yellowfin, skipjack and bigeye spawn several millions of eggs, several times each year). Spawning takes place over extensive areas and periods of time, where the larvae can always (statistically) find good conditions for their survival. In such wide scale conditions suitable for tropical spawning, recruitment will be limited by the carrying capacity of the large nursery areas or by environmental variability, rather than by the size of the spawning stock.

Temperate tunas, such as bluefin, may be quite different, as it is well shown by the decreasing trends of bluefin recruitment in the western Atlantic or in the southern bluefin tuna stock. The decrease in recruitment is probably a consequence of very reduced spawning stock size. The existence of spawning areas which are quite limited geographically and seasonally, and often very far from the feeding grounds, may be a dangerous component in the life cycle of those species.

The conclusion is that overfishing may have little or no negative effects on the fisheries catching juveniles of tropical tunas, but that this interaction may represent a real danger for temperate tuna. This danger of competition is of key importance, as the future of any stock relies on its recruitment level.

3.3 Medium Term Interactions

The medium term interaction addresses two different questions. These are (a) the extent fisheries catching small or medium size tunas affect fisheries catching large tunas, and (b) how longline and purse seine fisheries share the biomass of large tunas.

3.3.1 Interaction between fisheries catching small or medium size tunas, and fisheries catching large tunas

The first type of interaction (a) is directly a yield per recruit problem at the level of a stock (Ricker, 1958). It concerns all species which are exploited at a smaller size by surface fisheries and at a larger size by other fisheries. All tuna species with a large potential of growth (e.g., yellowfin, bigeye, bluefin, albacore) which can be caught at small sizes or at large sizes by various gears are potentially concerned by this type of interaction. The Atlantic Ocean offers good observations and relevant analysis since a good statistical database is available from this ocean through ICCAT. In the usual multi-gear yield per recruit models, a stock unit is accepted as an assessment and management unit. The fishing mortalities of each gear are first calculated under this hypothesis (following VPA and based on the catch by age for each gear), and then the potential for yield per recruit competitions between gears can easily be calculated.

This type of study, based primarily on the analysis of fishery data, is quite simple to conduct when detailed fishery data are available. However, they are usually heavily dependent of various parameters and conditions:

· The quality of catch statistics for the small fishes, especially the correct species composition of the catches. These “correct statistics” may often be difficult to obtain as it is difficult to estimate the actual quantities of small bigeye and yellowfin tunas caught by most fisheries. When the catch data are an order of magnitude incorrect (as it was often the case for juvenile bigeye tunas) all subsequent interaction analyses would be biased. Also, correct data on discards are often difficult to obtain but may be of significant importance in various fisheries, for instance with the development of fisheries associated with natural and artificial logs (where large amounts of very small tunas are often discarded).

· The exact nature of the stock structure and its year to year variability, including the existence of specific migratory routes between fisheries, intensive or low mixing rates between fishing areas, and effects of the environment on migrations. These biological parameters linked to the tuna migration and to their environment are partly described and categorised in Section 2. A more detailed analysis of these phenomena would require specific and ad hoc studies in each ocean and for each species.

· The natural mortalities M assumed for the juveniles. This factor may be of key importance in the interaction studies as the numbers of small individuals caught are often known (depending of the fishery statistics), but not the true numbers of small fish in the sea (recruitment). The yield per recruit calculations will be (most of the time) quite sensitive to this variability of M during the early life of the fishes. As most often M is a poorly known parameter, especially for juveniles which are entering in the fisheries, this problem may be of key importance to evaluate the potential interactions between juveniles and adults tunas fisheries.

Tunas in the same ecosystem at the same size should have similar natural mortalities.

This “reasonable” argument has been accepted since the early 1980s as a working hypothesis by the SCRS scientists. The hypothesis is that, within a plurispecific school of skipjack and juveniles bigeye and yellowfin tuna, the M of each species is very similar. Following this argument, it was accepted by ICCAT scientists that the M for juvenile bigeye was not M=0.4, as previously accepted, but M=0.8 (i.e., the M level accepted for skipjack tunas). The same rule was also applied for yellowfin tuna, and the juvenile M of this species increased from 0.6 to 0.8 during the first two years of their life. Those changes in the juvenile M has produced changes in the recruitment levels (increased levels), in the yield per recruit and in the potential competitions between fisheries catching juvenile and adults (decreased potential for interactions).

There are at least three general rules in the variability of the natural mortalities of any animal that should be kept in mind:

(1) The individual in their early life are more fragile and suffer a higher risk of natural mortality. This higher risk may for instance be due to a higher risk to be eaten by a predator (including by their own parents) or to die early by natural selection (“natural” death of the weakest individuals).

(2) Natural mortality tends to increase after sexual maturity for most animals. This feature has been noted in terrestrial fields where the natural mortality is more easy to measure (Gompertz, 1825).

(3) Most fishes living in warm equatorial or subtropical waters have a higher M than the individuals of similar size living in the cold or temperate waters (Pauly, 1980).

As a conclusion, the yield per recruit potential competition between fisheries catching juvenile tunas and fisheries catching adult tunas may be handled by the traditional yield per recruit analysis. However, the results obtained may be very sensitive to the various uncertainties in the data available, the stock structure and the profile of natural mortality by age.

3.3.2 How do longline and purse seine share the biomass of large tunas?

The relationship between fishing effort of purse seiners and longliners, and their respective catches of large fish, shows completely different patterns for yellowfin and bigeye tuna. The potential catches of the longline fishery to catch large yellowfin seems always limited, compared to the catches by the purse seine fishery; the latter uses an efficient and productive gear type (Figure 9). An opposite situation is observed for bigeye were the purse seine is an inefficient gear to catch large bigeye in surface layers, and the longline gear is very efficient to target large bigeye which generally inhabits deeper waters. This differential efficiency of each gear is probably related to the heterogeneity in the vertical distribution of each stock, yellowfin being basically a subsurface tuna, even at large sizes, whereas bigeye is most often a deep-swimming and temperate species at larger sizes.

Consequently, the performances of each fishery and the relative potential of interactions between them cannot be well estimated by the yield per recruit models. In fact this type of model will overestimate the efficiency of longliners on large yellowfin and overestimate the efficiency of purse seiners on large bigeye. However, there are still probably some potential real and direct interactions between the two gears on the adults when they are fishing in the same area: these interactions can be of the "urn" type interaction (cf. section 3.5), in which locally a tuna taken by one gear cannot be taken by the other gear, which gives an advantage to the more efficient gear.

3.4 Short Term Interactions

How do various gears fishing in the same time and area compete “locally” to exploit tuna resources? The answer to this question may be related to the local movement patterns of the tunas.

(a) The "urn" concept. When a fraction of stock is temporarily resident in an area, a simple “urn model”, assuming no significant input or output of fishes, can describe the local and short term interaction between the local fisheries. Tunas are clearly migratory; however, during limited periods of time, they can very often be considered as sedentary, and assumed to be a locally and temporally isolated fraction of stock. This type of seasonal concentrations are often observed for tunas in various circumstances and sites such as:

Spawning concentrations in offshore spawning areas (such as the yellowfin in the Atlantic and Indian Oceans).

Feeding concentrations in both inshore or offshore areas. Those type of feeding concentrations are often linked with the seasonality of the environmental conditions. For instance in the Atlantic Ocean, the area of Cape Lopez and Senegal offer good feeding opportunities seasonally, and this attracts a large biomass of tunas to the area each year.

Various islands and sea mounts ecologically attractive for tunas: it appears from both tagging and fishery data that the biomass of tunas available in the vicinity of islands is often locally resident during some months or even years, depending on the species and environmental variability.

When a fraction of stock is stable in this area of concentration (e.g., no emigration and no immigration), the local interaction problem becomes quite simple. During limited periods of time, it can be assumed locally that:

· The natural mortality becomes much less important.

· The fishing mortality rates can be easily calculated when the total catches are known, if there is a measure of local abundance (local CPUE for instance). If the local exploitation rate was high, a case often observed upon tuna concentrations fished by tropical purse seiners, the final local catch tends to be the initial biomass recruited in the area (very high local exploitation rate).

This type of short term and local competition is well shown in the Atlantic by the catch, effort and CPUE relationship of purse seiners: when the effort is low within small strata, the average CPUE is often high or very high. When the effort and catches are high in the same strata, the CPUE is more often moderate or low. In this case of high local effort, it is clear that there was probably, at least at the beginning of the period, a large biomass which is in contradiction with the low CPUE. The easiest explanation for the final low CPUE, is probably related to the urn concept: the interaction between a great number of boats on a limited local resource will increase the catch, but will reduce the CPUE, more quickly than a small fishing effort. The final CPUE of this large fleet, integrated over a monthly period, will then be an average low (instead of estimating the large initial biomass).

(b) The “pipeline” concept. When there is a permanent flow of tunas in an area, for instance a fishery taking place in a migratory route (similar to a salmon fishery during the migration towards spawning grounds), this interaction problem may be well treated as a “pipeline fishery”, with “upstream” fisheries catching migrating tunas which will not be available to the “downstream” fisheries.

4. A REVIEW OF VARIOUS POINTS RELATED TO INTERACTIONS BETWEEN TUNA FISHERIES

4.1 Tagging, Life Cycles and Spawning

It appears clearly that tagging remains the most direct way to estimate the transfer rates of tunas between fisheries and their potential for interactions. However, one should always be aware that movement information obtained from tagging is most often based on juvenile tunas (because of the low probability to recover tagged tunas after long durations for most species), and the data should not be extrapolated to the spawners. This precaution is because of the major physiological and behavioural changes developed by tunas during their life. A good example of this problem was shown recently in the Atlantic yellowfin fishery where, until 1982, a two-stocks hypothesis was assumed, based on the tagging of juveniles. A new “one-stock hypothesis” was developed recently developed from the tagging of medium-sized fish. Consequently the tagging of adult tunas, or other methods such as genetics or biochemistry of bones (which can potentially keep records of fishes past migration), should be developed in order to evaluate the movement patterns of adult tunas, especially in relation to spawning. These results may be fundamental to estimating most tuna interactions.

Figure 10. The “Urn model” concept of a temporarily isolated tuna resource (local concentration) heavily exploited by competing fishing units.

4.2 Interactions for Various Tuna Species

The potential for interactions and the nature of those potential interactions are highly dependent of the ecology and life cycle of each tuna species (especially the movement patterns). However, the interactions may be similar in all oceans for a given species. A three-group classification scheme, as developed in Section 2.3.1, with tropical species (skipjack and Auxis spp.), tropical-temperate species (yellowfin, bigeye and albacore) and temperate species (bluefin), could well be used as a useful basis to built ecologically similar groups with similar types of life cycles and migrations and similar potential for interactions. It would appear in a first review that the potential risk for interaction is probably more critical for the more temperate species, because of their longer duration of exploited life and larger potential of growth, smaller stock size and probably having more precise migration patterns and a more pronounced homing behaviour.

4.3 Genetic Heterogeneity of Stocks and Fishery Interactions

The genetic studies on tunas are still quite preliminary to make any definitive statement of the genetic structures of tuna populations. However, it can reasonably be assumed as a working hypothesis that the temperate or subtemperate species, may show a significant genetic heterogeneity, each genetic group showing a characteristic migration pattern toward a genetically sustainable unit (sub-population) linked with a strong homing behaviour (which is an evolutionary necessity for temperate tunas spawning in specific subtropical areas). In such an unusual (for tunas) but realistic hypothesis, the migratory routes of specific fractions of stocks would be of major importance for both the stock assessment and stock management, and more specifically for the potential interactions between fisheries. The studies of the genetic heterogeneity within the tuna populations and stocks should be developed to better understand the various types of fishery interaction.

4.4 Changes at the Stock’s Frontiers: Interactions or Environment Effects?

Tuna fisheries, especially those catching temperate tunas, often show striking and unexplained changes in relative abundances and catches. Such examples were observed for bluefin tuna in Canada, Norway and Sweden, and for bigeye in the Azores and Madeira. These changes of CPUE can be explained:

· by the environment, either by changes in the local environment (for instance waters too cold or too warm locally), or changes in the food locally available to the tunas, or by changes in the global environment for instance in remote places of the migratory routes;

· by interactions with upstream fisheries (“pipeline fisheries”), or by a simple decrease in the total stock abundance which often produces for pelagic species a reduction in the distribution area of the stock. In such a case of reduction in the distribution area, a moderate decrease of the stock total biomass can produce exaggerated and dramatic effects on the fisheries operating at the frontiers of the stock. This could create an unexpected and spectacular type of interaction between fisheries.

5. REVIEW OF SPECIFIC INTERACTIONS IN THE ATLANTIC

5.1 Small Tunas

The small tuna fisheries catches are reaching an average level of 127,000 mt in the Atlantic (i.e., 23% of the whole Atlantic tuna catches). Those catches are predominantly of Sarda spp. (25%), Scomberomorus spp. (26%) and Euthynnus spp. (17%). Few studies have been conducted in the Atlantic on these species caught by industrial and artisanal fisheries. For most species, the stock structure is completely unknown. However, because of the coastal nature of most of these species, and because of the short distance movements indicated by few recoveries, it would appear that the potential for interactions between fisheries catching small tunas would be limited to the regional or sub-regional levels. In this hypothesis the management of these resources should be conducted at a sub-regional level (e.g., off the Ivory Coast and Ghana, not the entire Atlantic), all the potential interactions between remote fisheries being assumed to be non-significant. In such a hypothesis, potential interaction may still exist between the various fisheries catching these small tuna species.

5.2 Skipjack Tuna

5.2.1 Fisheries

The skipjack fisheries (Figure 2) in the Atlantic caught an average of 147,000 mt per year during the period 1988-92. These fisheries were located predominantly in the eastern Atlantic (82%), but were also present in the western Atlantic (18%). The main gears catching skipjack tuna in the Atlantic were purse seiners (54% of the catches), and pole-and-line baitboats (45%).

5.2.2 Stock structure

A two-stock hypothesis is usually accepted by the SCRS in conducting skipjack stock assessments (but very little skipjack stock assessment was done during recent years by ICCAT scientists). This hypothesis was based on the geographical distribution of the various age groups caught in the Atlantic. No trans-Atlantic migration of skipjack has been demonstrated from the tagging results, but this result may not be significant, as all tagging was done in the eastern region, while few fisheries were active in the west. In the Atlantic, skipjack tunas show predominantly a large diffusion component in its movement patterns. This diffusion pattern is seasonally driven by the seasonal variability of the environment in the Atlantic, which produces large north-south movements of the skipjack biomass, especially in the eastern Atlantic.

5.2.3 Interactions

The SCRS scientists concluded that the Atlantic skipjack stock was, until now, moderately exploited. Most of the research conducted on the Atlantic skipjack was conducted in the early eighties, especially during the ICCAT International Skipjack Year (1981). The analyses were conducted on the CPUE by sizes of various fleets (purse seiners and baitboats) and on the tagging and recovery data done during this program. The general conclusion of these analyses concerning the potential interactions between the various fisheries was that little interactions should be expected with the fisheries in activity during this time. In fact, because of the low availability of both small (e.g., less than 45 cm) and large (e.g., greater than 60 cm) skipjack to the fisheries, and because of the relatively low exploitation rate of the stock, very little potential interaction should occur, at least between the large areas under study.

However, some interactions at a small geographical scale, between small adjacent fishing areas (for instance between the Ghana and Liberia areas, located at a distance of 1,000 nm, with fish moving freely between the two areas) could happen. These small scale interactions have not been quantitatively estimated. They could probably be classified as “urn type” interactions, with various fisheries competing temporarily to exploit a limited resource, which is stationary in a small geographical area.

5.3 Bigeye Tuna

5.3.1 Fisheries

The tuna fisheries in the Atlantic (Figure 2), caught an average 67,000 mt of bigeye per year during the period 1988-92. The fisheries are located predominantly in the eastern Atlantic between 30°N and 30°S). Three gears are primarily used to catch bigeye tuna in the Atlantic; these include longline (62%), purse seine (13%), and pole-and-line baitboats (12%). Longliners catch only large bigeye, while most baitboats and purse seiners catch bigeye of smaller sizes. An exception should be noted for the baitboat fisheries located in Madeira, Canary and Azore Islands, where they catch large bigeye.

5.3.2 Stock structure

A single-stock hypothesis is used by ICCAT scientists for the stock assessment of this species. The fisheries for juveniles are located in equatorial areas that are different from the areas of adult concentrations; the latter are located near the northern and southern tropics. This observation on distribution is easily explained by the physiology of the species. Young bigeye are physiologically a tropical species, living in surface warm waters. They become temperate and deep-swimming when mature, except during seasonal spawning when they return to tropical waters (Figures 6 and 7).

Very few scientific data are available from tagging of large fishes, or from other sources, to determine the real stock structure of bigeye tunas in the Atlantic. For instance, even the location and time of spawning are not well known. Furthermore, if the present hypothesis of one spawning zone and two major feeding areas (north and south temperate areas) is valid, it is still unclear when the adult bigeye do their spawning migrations toward equatorial areas. It is also unknown if the adults return to their original northern and southern feeding zones after spawning, or if they migrate randomly in any of those two directions. All of these stock structure characteristics may be of key importance in the study of potential interactions between fisheries.

5.3.3 Interactions

Even though juvenile bigeye tuna are quite difficult to recognise from juvenile yellowfin (e.g., Fonteneau, 1975), a systematic scientific sampling of the species composition of small tuna catches has been conducted in the Atlantic on a routine basis since 1979 (ICCAT, 1984). The sampling has provided reasonably good estimates of the numbers of small bigeye landed. The catch of small (less than 80 cm) bigeye taken in the Atlantic since 1955, the beginning of the industrial fisheries, are shown in Figure 11, together with the catch of large bigeye (more than 120 cm). Figure 12 shows the catches of large bigeye by gear (baitboat, purse seiners and longliners).

The CPUE of small bigeye caught in surface fisheries shows large fluctuations, but without a clear trend (see Figure 13a). The CPUE of large bigeye, as indicated by the longline GLM abundance index (see Figure 13b), shows a moderate and regular decline since the beginning of the fisheries. These indices suggest that recruitment was stable and that the decrease in adult biomass was moderate.

Figure 11. Catches of Atlantic bigeye tunas, by size category.

Small: less than 80 cm fork length
Medium: 80 to 120 cm.
Large: more than 120 cm.

Figure 12. Catches of large (+1.2m.) Atlantic bigeye tuna by gear

(BB=baitboat, PS=purse seine, LL=longline).

Figure 13. CPUE for bigeye tuna. (a, left) Surface fisheries--nominal CPUE of French purse seiners, and (b, right) longline fisheries--GLM indices from ICCAT, 1994b.

It is clear for bigeye that the adults caught by the purse seine fishery cannot have a significant effect on the longline fishery. This is due to the small purse seine catches of large bigeye (Figure 13) and to the differences in fishing areas of the two gears.

The present stock assessment analysis on bigeye concludes that the decline of the adult abundances can be explained by two factors:

· by the increased catch of the longliners, reducing directly the adult biomass;

· by the increased catches of juvenile; the large numbers of small bigeye ((3 to 5 millions of individuals yearly) taken by surface gears correspond to a moderate fishing mortality and has had only a moderate effect on adult stock size. This is shown by the results of the VPA, which give estimates of the recruitment in the stock and in the adult fishery. The bigeye recruitment seems to fluctuate without decreasing trend (Pereira, 1994). However, the decrease of this adult recruitment is clear, at a rate of approximately 20 % to 30 % between the “virgin” stock (e.g., the level of the stock during the early 1950s) and recent years, but is relatively moderate. It can be noticed also that present yield per recruit analysis indicates that the juvenile fishery affects the longline fishery, but does not significantly reduce the overall yields, at least at present levels of fishing mortalities (Pereira, 1994).

5.4.1 Fisheries

The tuna fisheries in the Atlantic, caught an average of 154,000 mt of yellowfin per year during the period 1988-92 (Figure 2). The fisheries are located predominantly in the eastern Atlantic (123,000 mt, i.e., 80% of the total catch), but also significantly in the western Atlantic (31,000 mt, i.e., 20%). Three gears are primarily catching yellowfin in the Atlantic: purse seiners (73% of the catches), pole-and-line baitboats (13%) and longliners (5%).

5.4.2 Stock structure

The migrations of yellowfin tunas between the eastern and western Atlantic are now assumed to be significant, and a single-stock hypothesis has been accepted by the SCRS scientists since June 1993 (Figure 14). This conclusion was based on the transatlantic recoveries of tagged fishes and on the catch and CPUE of fisheries.

Figure 14. The hypothetical stock structure of Atlantic yellowfin presently accepted by the SCRS in 1993.

5.4.3 Interactions

The yellowfin stock has been heavily exploited since the early 1980s; the stock producing an average catch more or less equal to the estimated MSY of the stock. The fishing zones of small yellowfin are quite similar to the adults, at least for purse seiners; however, small yellowfin are usually caught in more coastal waters than the adults. The adults of yellowfin taken by longliners also show a wider area of distribution, scattered between 10°N (reaching 50° N in the western Atlantic), and 10°S from America to Africa. The adult yellowfin taken by the purse seiners are predominantly caught in a restricted and small area, in the major spawning area of the eastern Atlantic (Figure 14).

The numbers of small (-70 cm), medium and large (+120 cm) yellowfin taken by Atlantic fisheries are given in Figure 15. Figure 16 shows the dramatic increase of small yellowfin in the catches during recent years by purse seiners, and the relative stability of the large yellowfin catches since 1975 (decreased catches by longline and increased catches by purse seiners).

The catch and CPUE of small yellowfin, (taken by baitboats and purse seiners), showed some year to year variability, but no definite trend, suggesting a relatively stable recruitment. The abundance trend of large yellowfin measured by the CPUE on large yellowfin by longliners and purse seiners showed for recent years a moderate decline. However this decline is quite difficult to evaluate precisely for both longliners and purse seiners. The purse seiners in recent years have increased their fishing power due to changes in nets, the use of bird radar and other computerised facilities. The practical effect on the catch rates of those changes are difficult to evaluate precisely, but the recent abundances of yellowfin may be overestimated. The change of target species from yellowfin to bigeye occurred in recent years with a shift in fishing zones and increased fishing depth of the longline (the densities of bigeye being higher at greater depths, densities of yellowfin being lower). Consequently, it is quite difficult to calculate a CPUE index for longliners that is representative of yellowfin abundance.

Figure 15. Catches of large (+120 cm) Atlantic yellowfin tunas by gear

(BB=baitboat, PS=purse seine, LL=longline).

Figure 16. Catch of Atlantic yellowfin tunas, by size category.

The decline of the adult CPUE corresponds to a decline of the adult abundance. The interpretation of the VPA results (ICCAT, 1994b) can be that this decline of adult stock is a consequence of two factors: (a) the development of fisheries for juveniles which reduces the recruitment to the adult stock, and (b) the increased effort toward, and catches of, adults which directly reduces the adult biomass.

Present analysis suggests that the first effect--interaction between fisheries for juvenile and adult yellowfin--would be significant. This is shown by the decline of the adult recruitment estimated by VPA at approximately 50% between the early period of the fisheries and the present fisheries; the decline due to the increased catch of juveniles. However, subsequent yield per recruit analysis shows that the juvenile fishery does not seriously decrease the yield per recruit of the overall fishery because the juvenile catch is only slightly less (in weight) than the subsequent theoretical loss of the adult fisheries (at least with a high M estimated for the juveniles) (ICCAT, 1994b).

5.5 Bluefin Tuna

5.5.1 Fisheries and stocks

The bluefin fisheries in the Atlantic, caught an average of 26,800 mt per year during the average period 1988-92. The fisheries are located predominantly in the eastern Atlantic (6,400 mt) and Mediterranean Sea (17,400 mt), but also some bluefin are caught in the western Atlantic (3,000 mt). Various gears are used to catch bluefin tuna in the Atlantic, but purse seiners (45% of the catches) and longliners (17%) are the most important.

5.5.2 Stock structure

The migrations of bluefin tunas between the eastern and western Atlantic have been well shown by tagging and transatlantic recoveries. Nevertheless, until November 1993, a two stocks hypothesis assuming no significant east/west mixing and no potential interactions between the eastern and western Atlantic fisheries was assumed by the SCRS scientists. However, a mixing stock hypothesis seems now to be better accepted by various scientists, following the recommendations developed by experts from the USA (NRC group of experts) in 1994 (Figure 17). This stock structure hypothesis is clearly of key importance in the stock assessment and management of the Atlantic bluefin stock, even if the mixing rates between east and west bluefin populations are low (about 3% yearly?). The importance of the mixing and the potential interactions between stocks and fisheries located in both sides of the Atlantic are becoming major issues in the ICCAT stock assessment and management.

The condition of the bluefin stock in the western Atlantic appear to be consistently poor since the early 1980s, despite the establishment of small catch quotas in this area to increase population size. This stock has shown during recent years a critical decreasing trend of recruits and spawning biomass. On the contrary, the bluefin stock in the eastern Atlantic and the Mediterranean Sea seem to be in better condition; the recent increases in catches following the increasing effort have been sustained. Still this stock is showing a spectacular decrease of its spawning biomass, and its exploitation rate seems to be presently very high for most age groups. However, recruitment levels in the eastern Atlantic remain stable and high, being estimated at millions of fish yearly.

Figure 17. The hypothetical stock structure of Atlantic northern bluefin being discussed by the SCRS.

5.5.3 Interactions

The Atlantic bluefin tuna is a species with a large potential growth (the giant bluefin may exceed 500 kg) and low natural mortality (M estimated at 0.2, i.e., 18% of the population dying yearly of natural causes). A logical consequence is that there is a large potential of competition between the fisheries catching small bluefin and those catching large individual. This quite obvious yield per recruit competition was at the origin of the bluefin size limit of 6.4 kg imposed by ICCAT since 1975. The Atlantic bluefin may also offer various interesting other types of interactions between fisheries. During past years this species has shown that some local fisheries that have been active for years vanished suddenly without any clear scientific explanation; e.g., Norway, Morocco and Brazil provide good examples of collapsed fisheries. It is not clear if those negative changes were linked with an environmental variability or to a competition between fisheries. They may be due, for instance, to changes in the local environment (for instance waters getting too cold or to warm for the tunas), or to a decrease in food locally available to the tunas, e.g., the overexploitation of the local small pelagic resources may also explain why bluefin tunas disappeared from some areas. They also may be due to a possible heterogeneity of the stocks as explained earlier.

In the stock assessment conducted by ICCAT scientists until 1993, no potential interaction could exist between the eastern and western Atlantic fisheries since a two-stock hypothesis was assumed. In the “1994 mixing stock hypothesis” (proposed by the NRC experts and accepted by SCRS scientists), there is a potential interaction between the eastern and western Atlantic bluefin stock and fisheries. Because a huge population is recruited in the eastern side and a very small one in the west, even a low mixing rate is sufficient to bring significant numbers of bluefin into western areas. Any error or inter-annual variability in this east-to-west mixing rate may change dramatically the absolute numbers of fishes in the western component (this variability is thought to be important based on bluefin tag recovery data). This problem is now being analysed by the SCRS scientists (primarily from tagging recovery data and modeling).

5.5.4 The standard “homogeneous” stock assessment

In the standard stock assessment, each stock is implicitly assumed to be a homogeneous unit; each individual fish shows the same behaviour as other fishes from the same cohort. So, in this case the stock (i.e., the management unit) corresponds to a single homogeneous population (genetic unit). Consequently, the decreasing biomass trends may often be reversed. If the stock recruitment relationship allows for a compensation, the stock will slowly return to its original levels under a reduced (or null) fishing mortality (stability of the MSY).

5.5.5 Bluefin tuna: a genetically heterogeneous stock?

The hypothesis of a genetic heterogeneity may well be accepted for bluefin tuna, because of the possible homing behaviour of the species and because of the long life expectation (each fish will spawn for about 10 to 15 years which allow some stability in time for such units). In this hypothesis, bluefin could be like salmon or marine turtles; thus, instead of having a single homogeneous population, there could be various self sustaining and independent genetic units or multiple sub-populations. Each unit would spawn in a precise time and area strata, and would conduct an exact migration pattern, characteristic of each genetic group; at the age of first spawning, this group of fishes would return to meet and spawn in a specific site and time with individuals belonging to the same genetic unit.

If this situation exists, genetic units may be extinguished. The loss of the genetic unit could result from repeated adverse environmental conditions in the migratory routes. Local and repeated overfishing in a specific point of the migratory route may also produce an extinction of this original genetic unit. The loss of the genetic unit could also result from environmental conditions becoming adverse in the geographically fixed spawning strata. This lack of flexibility in the choice of the spawning strata is probably a major biological necessity for all homing species. It may have strong evolutionary advantages in the long term, as shown by Cury (1994), but it may be fatal in the short run for a given genetic unit. In this case, the stock may never recover to its original levels, because of the loss of its genetic diversity. The MSY would then be decreased following the overexploitation or the changes in the environment (with a corresponding decrease of the optimal fishing effort).

If the absence of bluefin tuna presently observed in some of their traditional feeding grounds corresponds to this "cascading extinction of various sub-populations", various bluefin stocks may be following this scheme of heterogeneous genetic units. This loss of genetic diversity and of potential biomass of the stock may be irreversible in the short term. In the longer and evolutionary terms, the adaptive potential of migrating and spawning strays, would probably allow a stock recovery under a different biological and geographical pattern (for instance creating new spawning areas, other migratory routes and increased potential catches for fisheries). Such a case may have been developed in the Mediterranean Sea, when this sea was colonised by spawning (and homing) bluefin tunas, which possibly strayed from the western Atlantic.

5.6 Albacore

5.6.1 Fisheries

The albacore fisheries in the Atlantic caught an average 63,000 tons per year during the average period 1988-92 (Figure 2). They were located predominantly in the North Atlantic (31,300 mt), South Atlantic (27,900 mt), and Mediterranean Sea (2,900 mt). This present level of catches is significantly lower than the 80,000 tons yearly catches observed until the 1960s and early 1970s. Three gears traditionally catch albacore: longline (50%), trolling and baitboat. However, during recent years a significant increase of the catches by new gears, midwater trawling and drift gillnets, were observed in the northeastern Atlantic (15% of the North Atlantic catches).

5.6.2 Stock structure

The hypothesis of three independent albacore stocks, in the North and South Atlantic and the Mediterranean Sea, has always been accepted by the SCRS. This hypothesis is well supported by the absence of north-to-south migration of tagged albacore and by the prevailing low densities of albacore in the inter-tropical Atlantic.

5.6.3 Interactions

The present status of albacore stock in the North Atlantic is being discussed by ICCAT scientists. This stock appears to be presently under-exploited (as indicated by both the global and analytical models), because of the decrease of fishing effort by surface and longline fisheries (for the obvious economic reasons of a lack of efficient fishing gear and the low value of the species in most tuna markets). The South Atlantic stock seems to be fully exploited, with longlines being the major gear exploiting this stock.

The problem of potential interactions in the North Atlantic albacore fisheries has not been examined in detail. However, various types of interactions can be shown for this stock, for instance:

· The yield per recruit interaction between the surface fisheries and the longline fisheries. The surface fisheries catch juvenile albacore (ages two to four), below the optimal size (in term of yield per recruit), thus reducing recruitment to the longline fishery (ICCAT, 1994).

· The local interactions between various surface gears operating in the Bay of Biscay; the fisheries include trollers, midwater trawlers, pole-and-line, baitboats and driftnets.

· either a simple “urn competition”, where the more efficient gear catches part of the local biomass before the other less efficient gear, thus reducing the CPUE,

· or possibly a decrease of the efficiency of the “traditional” gear due to the introduction of the new gear. In such an interaction, the fishes could for instance be frightened by the drift gillnet, or their movement patterns could be modified, the consequence being a lower catchability for the traditional gear (the total local biomass being unchanged). However, it was not possible until now to confirm this hypothesis by an ad hoc observers program.

5.7.1 Fisheries

The swordfish fisheries in the Atlantic Ocean have shown a spectacular increase in their activities in recent years, with the catches increasing from less than 12,000 tons per year during the 1960s and 1970s, to an average 44,000 tons during each of the last five years (Figure 2). The major gear catching swordfish was longline (78%). During the same recent period, the catch was geographically distributed as following: 35% in the North Atlantic, 30% in the South Atlantic and 35 % in the Mediterranean Sea.

5.7.2 Stock structure

The present SCRS analyses has worked under the hypothesis of three independent stocks of swordfish in three areas, with no potential interactions. This hypothesis was based on the analysis of fishery data and tag and recovery data. Unfortunately tag release and recovery rates are low for swordfish. In fact, the fishing zones of swordfish in the Atlantic show a north-south continuum of the high abundance and catches, which suggests that the 5° N limit used by ICCAT scientists may not be a real biological frontier between a northern and a southern so-called “stocks”. This is unlike the clear case for albacore. This hypothetical stock structure is in fact highly problematic in the absence of strong scientific data to support it.

5.7.3 Interactions

The Atlantic swordfish is considered to be presently heavily exploited, at least in the North Atlantic. The yield per recruit analysis presently conducted on swordfish indicates clearly that there is a serious potential for competition between the fisheries targeting small individuals and the fisheries catching larger fish (at least presently under high exploitation rates). The ICCAT size limit adopted on North Atlantic swordfish since 1990 was based on this yield per recruit analysis.

The exact level of this yield per recruit competition is highly questionable and dependent upon a number of uncertainties, e.g., the recruitment trends and actual present exploitation rate of the stock(s). The present stock assessment of swordfish may provide a good example of the scientific errors done with VPA when the fishing patterns and exploited fishing zones are changing. Such changes, primarily the increasing size of the fishing zone, may have produced an increased recruitment which was not a biological reality (at the level of the true Atlantic swordfish population). It is also quite obvious that swordfish provide a good example of a tuna like species which is scattered ocean-wide (and in a wide range of depth, including very deep waters), when only the areas with the highest densities are exploited, and when only a fraction of the stock is presently exploited (as suggested by Figure 18). Consequently this species may still have a significant but unknown cryptic biomass in the Atlantic, thus the present fishing mortalities, exploitation rates and potential interactions may have been overestimated. The present stability of the total swordfish catches in the North Atlantic since 1989, taken by a stable fishing effort (ICCAT, 1994a), may also be an indication that the swordfish global exploitation rate may have been overestimated by the SCRS (as the very high exploitation rate presently estimated should produced soon a decrease of the catches).

5.8 Other Billfishes

5.8.1 Fisheries

Various billfishes are exploited in the Atlantic Ocean. Most billfish are caught as a bycatch in industrial fisheries that target tunas. Three major species are taken in these fisheries: sailfish (2,100 mt annually), blue marlin (2,900 mt) and white marlin (1,300 mt). These fish are often discarded at sea and not recorded in the ICCAT statistics. Another significant part of the billfish catch is taken by the artisanal fisheries of Africa for local consumption and by sport fisheries, the major sport fisheries being located in the western Atlantic.

5.8.2 Stock structure

Very few studies are available on the stock structure of most billfishes. However, it is well shown by tagging that some species, e.g., blue marlin can undergo extensive trans-Atlantic migrations, thus allowing for a potential interaction between western and eastern Atlantic fisheries. A single Atlantic stock hypothesis was then used on this stock by the SCRS. Other species such as sailfishes show a limited potential for migrations (well shown by tagging and recoveries, that movement is always observed within a relatively short distance), and the stocks are probably limited to a regional scale. The stock assessment for sailfish is conducted under a two stocks hypothesis (eastern and western Atlantic).

5.8.3 Interactions

Very few studies of interactions between billfishes fisheries have been conducted, primarily because very few data were available. An interesting potential interaction case for billfishes is the local competition between commercial and recreational fisheries. Various countries, e.g., such as Venezuela and USA have developed policies giving priority to the development of sport fisheries. Governments have consequently put serious restriction to the access of EEZ waters where sport fisheries are active. The goal of those “intuitive” management measure was to reduce the local competition between sport and commercial fisheries, and to leave the entire local biomass available to the sport fishery (“urn” concept).

Unfortunately, no ad hoc scientific analyses were submitted to ICCAT, until now, to evaluate the practical effects of those regulations targeting on “local interactions”. Such local interaction studies in which a competing fishery can be suppressed and the subsequent effects on the remaining fishery, would be of major interest to be studied in detail.

Figure 18. Areas of swordfish distribution in the Atlantic as shown by the monthly numbers of swordfishes caught (by 5 degree squares) by Japanese longliners between 1956 and 1993, compared to the present major fishing zones of fisheries targeting swordfish. The positions of catches are drawn randomly within each 5 degree square; only the monthly catches with more than 50 swordfish caught are represented. This figure indicates that swordfish are significantly present in a large area between 50°N and 40°S with very low or no fishing activities.

6. CONCLUSIONS

The analyses of interactions between tuna fisheries were developed explicitly during the early- and mid-1980s under the coordination of the FAO Fishery Department. These studies have been developed quite intensively during recent years especially in the western Pacific Ocean. This work has shown, as did similar studies conducted in other oceans, the real complexity of this problem, including the wide distribution of most tuna stocks (geographically and vertically), the complexity of their movement patterns (which are variable between species and for each species between ages), and other multiple and heterogeneous factors which are often very difficult to integrate. Three types of research have been successfully developed in recent years:

1) A first progress has been obtained for various stocks with the improvement of fishery statistics (for instance better coverage of the statistics, better size composition of the catches and better species identification), allowing an improvement of the basic yield per recruit analysis.

2) The recent tagging programs successfully conducted on various stocks have provided more direct evidence concerning the tuna movements of various species(for instance in the western Pacific), a key element in the study for potential interactions.

3) The integrated modeling of fishery data and tag-recovery data has also shown very significant progress during recent years.

These studies need to cover the relationship between tunas and their environment, the behaviour and genetics of the tunas, their age specific changes of physiology and thermoregulation, their growth, the complexity and variability of their age specific movements patterns (from diffusion to advection), their age specific natural mortality, etc.. This list of important problems to consider in the analysis of potential interactions could cover most of the biological fields.

Further research, in both field of data collection and data analysis, are now necessary to evaluate better the real potential for interactions between various present or future tuna fisheries.

7. ACKNOWLEDGEMENTS

This paper was originally submitted at the FAO-sponsored meeting on interactions between tuna fisheries held in Shimizu, Japan (23-31 January 1995). Many changes and corrections were done to the original text of the conference to give its present form. The editors have provided invaluable help in improving this text, and an unknown referee also provided interesting views and criticism of the original text. We express our deep and sincere thanks to both for the interest and time invested in the revision of this paper.

8. REFERENCES CITED

Beverton, R.J.H., and S.J. Holt. 1957. On the dynamics of exploited fish populations. Fishery Investigations (Series 2) 19: 1-533.

Brill, R.W. 1994. A review of temperature and oxygen tolerance studies of tunas pertinent to fisheries oceanography, movement models and stock assessment. Fish. Oceanogr. 3(3): 204-216.

Boggs, C. 1994. Methods for analysing interactions of limited-range fisheries: Hawaii’s pelagic fisheries. In: Shomura, R.S., J. Majkowski and S. Langi (eds.). Interactions of Pacific tuna fisheries. Proceedings of the First FAO Expert Consultation on Pacific Tuna Fisheries, 3-11 December 1991, Noumea, New Caledonia. Vol. 1: Summary report and papers on interaction. FAO Fish. Tech. Pap. (336/1): 44-91.

Cayré, P. 1990. Les migrations: un comportement déclenché par l’environnement. Rec. Doc. Scient. ICCAT 32(1): 158-168.

Cayré, P., F.X. Bard and T. Diouf. 1988. Les migrations des thonidés de l'Atlantique. In: Fonteneau, A. and Marcille (eds.). Dans Ressources, pêche et biologie des thonidés tropicaux de l’Atlantique. FAO Tech. Doc. 292: 111-156.

Cury, P. 1994. Obstinate nature: an ecology of individuals. Thoughts on reproductive behaviour and biodiversity. Can. J. Fish. Aquat. Sci. 51: 1664-1673.

Deriso, R.B., R.G. Punsly and W.H. Bayliff. 1991. A Markov movement model of yellowfin tuna in the eastern Pacific Ocean and some analysis for international management. Fish. Res. 11: 375-395.

Die, D.J., V.R. Restrepo and W.W. Fox. 1990. Equilibrium production models that incorporate fished area. Trans. Am. Fish. Soc. 119: 445-454.

Finch, C.E. 1990. Longevity, Senescence and the Genome. University of Chicago Press. 922 p.

Fonteneau, A. 1975. Note sur les problèmes d'identification du patudo dans les statistiques de pêche. Rec. Doc. Scient. ICCAT 5(1): 168-171.

Fonteneau, A. 1991. Interactions entre pêcheries et gestion des ressources thonières dans les zones économiques exclusives. In: Le Gall, J.Y., X. Dereviers and C. Roger (eds.). Actes de la conférence thonière régionale pour l’Océan Indien sud ouest, Antananarivo, Madagascar. Colloques et séminaires ORSTOM.

Gompertz, B. 1825. On the nature of the function expressive of the law of human mortality and a new mode of determining the value of life contingencies. Phil. Trans. Roy. Soc. (London) 115: 513-585.

Harden Jones, F.R. 1981. Fish migration: strategy and tactics. In: Aidlay, D.J. (ed.). Animal migration. Cambridge University Press, Cambridge, England: 139-165.

Hilborn, R., and J. Sibert. 1986. Is international management of tunas necessary? South Pac. Comm. Newletter 38: 31-40.

ICCAT. 1984. Report of the working group on juvenile tunas (text and appendices). Coll. Vol. Scientific Doc., Vol. XXI(1).

ICCAT. 1994a. Report for biennial period, 1992-93, part II (1993).

ICCAT. 1994b. Report of the ICCAT working group to evaluate Atlantic yellowfin tuna (Spain, 1993). Coll. Vol. Scientific Doc., Vol. XLII (2)

Kleiber, P., and B. Baker. 1987. Assessment of interaction between North Pacific albacore, Thunnus alalunga, fisheries by use of a simulation model. Fish. Bull. 85(4): 703-711.

Laloe, F. 1989. Un modèle global avec quantité de biomasse inaccessible dépendant de la surface de pêche. Application aux données de la pêche d’albacores (Thunnus albacares) de l’Atlantique est. Aquat. Living Resour. 2: 231-239.

Lenarz, W.H. 1971. Yield per recruit of Atlantic yellowfin tuna for multigear fisheries. ICCAT SCRS/71/27: 17p..

Pauly, D. 1980. On the interrelationship between natural mortality, growth parameters and mean temperature in 175 fish stocks. J. Cons. Int. Explor. Mer 39(2): 175-192.

Pereira, J. 1994. Analyse de l’état du stock de patudo de l’Atlantique. Rec. Doc. Scient. ICCAT. Vol. XLII: 279-289.

Ricker, W.E. 1958. Handbook of computations for biological statistics of fish populations. Bull. Fish. Res. Board Can. 119: 300 p.

Ricker, W.E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191: 382 p.

Sharp, G.D. 1978. Behavioral and physiological properties of tunas and their effects on vulnerability to fishing gear. In: Sharp, G.D., and A. Dizon (eds.). The physiological ecology of tunas. Academic Press, New York: 397-449.

Shomura, R.S., J. Majkowski and S. Langi (eds.). 1994. Interactions of Pacific tuna fisheries. Proceedings of the First FAO Expert Consultation on Pacific Tuna Fisheries, 3-11 December 1991, Noumea, New Caledonia. Vol. 1: Summary report and papers on interaction. FAO Fish. Tech. Pap. (336/1): 326 p.