Page 14 - 10.1.1.498.8230

1312 C. Ravier and J.-M. Fromentin
principal sources of variability (Fromentin et al., 1998;
Bjørnstad et al., 1999).
Long-term fluctuations in fish abundance have been
demonstrated for several Atlantic fish populations (see
Cushing and Dickson, 1976; Cushing, 1982; Alheit and
Hagen, 1997). They have been commonly related to
three main factors:
(1) human activity, mainly through overexploitation
and pollution of spawning and nursery areas,
(2) environmental changes,
(3) biotic processes, such as predation, cannibalism
and competition.
Overexploitation is a well known feature of many
North Atlantic fish populations (Myers et al., 1996;
Cook et al., 1997). However, it is unlikely to explain the
100–120 years cycle of variations in bluefin tuna abun-
dance because, first, exploitation cannot lead to cycles of
about the same length (more than a century) over
different periods and in different locations, second, traps
are passive gear, and third, the number of traps
remained more or less constant over the period analysed,
so effort is likely to have been constant. In the case of
bluefin tuna, therefore, it is likely that factors (2) and/or of the resonant effects. Furthermore, Fromentin and
(3) were more influential. Biotic and environmental
factors can effect long-term variations in bluefin
tuna trap catches, through enhancing or impacting
recruitment. Environmental events mainly affect fish
recruitment through changes in:
sea temperature, which influences the daily develop-
ment and mortality of the eggs and fish larvae (Pepin,
1991; Ottersen and Sundby, 1995),
food availability, i.e. the match-mismatch hypothesis
(Hjort, 1926; May, 1974; Cushing, 1990) and/or the
impact of the wind and current at a large (dispersal of
fish eggs and larvae, Hjort, 1926; Lasker, 1975) and
small scale (turbulence playing a role in prey encoun-
ter rates; MacKenzie and Leggett, 1991; MacKenzie
et al., 1994; Kiørboe and MacKenzie, 1995).
Biotic processes, such as predation, cannibalism, and
competition resulting from food and habitat limitations,
can also generate cycles and long-term fluctuations in
fish stocks, through density-dependent-mortality/growth
and resonant effects (Caley et al., 1996; Knell, 1998;
Bjørnstad et al., 1999; Fromentin et al., 2000b).
The environment certainly influences migration
patterns of fish populations (e.g. the West Greenland
cod stock; Cushing, 1982). It could therefore be argued
that the 100–120 years cycle in catches of bluefin tuna by
the trap fishery may have been the result of a switch
between Mediterranean and West Atlantic spawning
sites attributable to changes in oceanographic conditions
and/or food availability. Past research and mark-
recapture studies indicated that bluefin tuna migrate
annually to well-defined spawning areas, i.e. they con-
duct established ‘‘homing behaviour’’ (Cury et al.,
1998). Moreover, Polovina (1996) demonstrated decadal
variations in the trans-Pacific migration of northern
bluefin tuna in relation to climate-induced change in
prey abundance, but also stressed that the spawning
areas never changed. Such arguments tend to dismiss a
hypothesis of changes in migration patterns, but it must
be kept in mind that most of studies cited were per-
formed over a short period and covered recent trends
only.
Our main hypothesis is therefore based on the old, but
still pertinent, concept of Hjort (1914, 1926), according
to which the 100–120 years cycle in bluefin tuna abun-
dance would result from annual fluctuations in recruit-
ment. Eastern bluefin tuna are widespread and only
reproduce in a small spatial and temporal window
(around the Balearic islands and in the south Tyrrhenian
Sea from mid-May to June, Mather et al., 1995). There-
fore, recruitment of bluefin tuna is more likely to vary
from year to year than recruitment of tropical tuna
populations, which reproduce over large areas and in
more stable waters (Fromentin and Restrepo, 2001).
This feature is of particular interest, because Bjørnstad
et al. (1999) demonstrated that short-term variability in
recruitment of fish combined with cannibalism between
juveniles may induce long-term fluctuations because
Fonteneau (2001) showed that noise in recruitment
combined with a large number of classes of spawners
could also lead to long-term variations in spawning
stock biomass and yields, as well as to regular cycles,
depending on the lifespan of the species. These last
theoretical findings appear consistent with the present
results, which show a 15–30 years cycle concomitant to
the bluefin tuna lifespan. This feature could further
provide an hypothesis to explain the collapse of the
Nordic fishery for bluefin tuna in the mid-1960s (Tiews,
1978; Marsac, 1998) and its reappearance after 30 years
(Olafsdottir and Ingimundardottir, 2000).
Fisheries implications
Atlantic bluefin tuna are mainly managed by quotas
and size limits (ICCAT, 2001). In the long term, the
Convention upheld by the International Commission for
the Conservation of Atlantic Tunas (ICCAT) as well as
the precautionary approach imply that Atlantic tuna
stocks should be managed by strategies based on
maximum sustainable yield, i.e. based on models that
often assume that the population is at an equilibrium or
steady state (Hilborn and Walters, 1992). The current
results have shown annual and long-term variations in
abundance of bluefin tuna, so the concept of maximum
sustainable yield may be irrelevant for the eastern
Atlantic population. The temperate life history traits of
the species (slow growth, late age at maturity, short