Predation and Cannibalism in Pelagic Predators: Implications for the
Dynamics, Assessment and Management of Pacific Tuna Populations
reports (PDF): FY
2010, FY 2009, FY
2008, FY 2007
The proposed work seeks to formally assess the hypothesis that the production
of economically important tuna stocks has been enhanced by the depletion
of large-bodied predators. If this hypothesis is correct, then fisheries
policies must consider the direct and indirect effects of fishing and
their ultimate impacts on management measures and objectives. For instance,
if the production of the highly-valued tuna fisheries has indeed been
enhanced through the depletion of sharks, marlins, and large-bodied
tunas, then the objective of maximizing fisheries yield and profits
is best met through continued suppression of large-bodied predator abundance.
If this hypothesis is deemed implausible, then recent claims that tuna
stocks have declined to ~10% of their pre-fishing levels are not consistent
with general principles of population biology and compensatory responses
in food webs.
Researchers propose to evaluate the hypothesis that tuna productivity
has been enhanced by predator depletion through an approach that combines
a research synthesis of data on the feeding habits of large predators
with simulation models of tuna populations. These models will be used
to analyze alternative fishing strategies by first defining plausible
representations of intra-guild predation and cannibalism, and then exploring
the implications of these interactions on management strategies developed
to meet fisheries objectives. The specific objectives are to:
(1) Quantify the magnitude of feeding on skipjack, yellowfin and bigeye
tunas by conspecifics and heterospecifics throughout the Pacific Ocean
(2) Explore how that feeding varies temporally and regionally
(3) Identify tuna life history stages vulnerable to each predator by
constructing prey-size-spectra for each predator species
(4) Couple age-based modeling approaches with bioenergetics models to
estimate predation mortality for each stage of each tuna species
(5) Explore the implications of predation and possible predator depletion
for policy-relevant reference points.
Research synthesis of pelagic food habits data
The goal is to describe the sources, magnitude, and variability in predation
on bigeye, yellowfin and skipjack tunas, and to identify the life history
stages when predation impacts are most important. Project researchers
will use the existing large body of research conducted over the past
half-century on the food habits of tunas, sharks and billfishes via
a formal research synthesis (Cooper and Hedges 1994). It is anticipated
that the existing body of research potentially suffers from confounding
effects caused by differences in sampling methodology, data reporting
and analysis across studies, as well as the absence of simultaneous
sampling in different ocean regions. Yet, the advantage of the research
synthesis approach is that by thoughtful statistical treatment of these
data, which attempts to identify and remove confounding factors, it
is possible to derive benefits from data already collected a trivial
fraction of the cost of conducting novel research (Zeller et al. 2005).
Moreover, because the literature on food habits studies date back to
the 1950's, it is possible to test for long-term shifts in food habits
that might have accompanied shifts in food web structure caused by fishing.
The focus will be on the volumetric or mass contribution of skipjack,
yellowfin, and bigeye tuna in the diets of potential predators. These
predators include marlins (blue, black, white, striped), pelagic sharks
(bigeye thresher, oceanic whitetip, silky, white, shortfin mako, blue)
and tunas (skipjack, yellowfin, bigeye, albacore, bluefin).
The first stage of the analysis is to collect and digitize all information
that is available on the food habits of these fishes. Project researchers
have already identified 24 data sources. The second stage is to code
each data source based on the breadth and detail of available information.
Researchers propose an initial organizational framework that recognizes
the hierarchical nature of the available data and permits a standardized
approach to evaluate each data source.
Once data are entered and coded, two separate types of analyses are
proposed. The first is to describe the mean contribution of each tuna
(by life history stage) to the diets of each predator species, and to
explore how this contribution varies by region, season, and across time
periods. Classification and Regression Trees (CART) will be used as
an initial exploratory tool to partition the variance in predation attributable
to these effects. Subsequent analysis may include, but not be limited
to, generalized additive models (GAMs) (Hastie and Tibshirani 1990)
and generalized linear mixed models (GLMs).
The second analysis will use the detailed data available in contemporary
studies to describe the prey-size spectrum of each predator species.
The prey size spectrum describes the relationship between predator body
sizes and the size range of food items eaten (Cohen et al. 1993). This
relationship is thought to reflect a combination of morphological constraints
on feeding (Magnuson and Heitz 1971) as well as foraging decisions presumably
made to maximize the trade-offs between energy gain and handling time
Researchers propose to develop simulation models of skipjack, yellowfin,
and bigeye tuna populations to assess the range of impacts that tuna,
shark, and marlin predation may have on stock productivity and on the
values of key fisheries reference points. A simulation modeling framework
over an estimation modeling framework was chosen for several reasons.
Foremost is the immense data requirements of estimation models that
attempt to simultaneously estimate tuna and predator stock dynamics
and their interaction terms, all in a size-specific manner. Such a modeling
effort may indeed prove ultimately useful, but a simulation modeling
approach such as the proposee here can be used to quickly screen alternative
plausible representations of tuna dynamics, and thereby reveal whether
a more intensive estimation modeling exercise is warranted.
This modeling approach combines two well developed quantitative tools
- age-based population models and bioenergetics modeling. The former
provides an ideal framework for this work, because it contains the minimum
degree of model complexity (i.e., age- and size-structure) needed to
address the hypothesis, and it also uses the same parameters and assumptions
as those used in tuna stock-assessment models. This therefore permits
an efficient and careful transfer of information about tuna demographic
rates, and it allows presentation of the simulation modeling results
in a context already familiar to stock assessment modelers. Bioenergetics
modeling represents the best-developed tool for quantifying predator
demand, having a long history in fisheries ecology (Stewart et al. 1981),
and is increasingly being applied to explore large-scale shifts in marine
communities (Essington et al. 2002, Williams et al. 2004).
Once plausible parameterizations of each species' population model is
complete, researchers will perform Monte-Carlo simulation runs of the
models where they estimate key biological and policy-relevant parameters
for multiple possible model parameterizations. Because the historical
status of tuna predators is a key uncertainty, they will avoid assigning
probabilities to either end of the continuum, but instead consider alternative
scenarios of "high historical biomass" and "low historical
biomass". Policy-relevant variables include FMSY, BMSY, and the
optimal allocation of effort across gears and purse-seine methods. Biologically
relevant variables include maximum reproductive rate at low and high
predator abundance, and ranges of biologically plausible historical
and contemporary tuna biomasses.
1 funding for this 2-year project estimated to be available mid-2006.
E. L. 1976. Optimal Foraging: attack strategy of a mantid. American
Cohen, J. E., S. L. Pimm, P. Yodzis, and J. Saldaña. 1993. Body
sizes of animal predators and animal prey in food webs. Journal of Animal
H., and L. V. Hedges, editors. 1994. The handbook of research synthesis.
Russell Sage Foundation, New York, NY.
T. E., D. E. Schindler, R. J. Olson, J. F. Kitchell, C. Boggs, and R.
Hilborn. 2002. Alternative fisheries and the predation rate of yellowfin
tuna in the Eastern Pacific Ocean. Ecological Applications 12:724-734.
T., and R. Tibshirani. 1990. Generalized Additive Models. Chapman Hall,
J. J., and J. G. Heitz. 1971. Gill raker apparatus and food selectivity
among mackerels, tunas and dolphins. Fisheries Bulletin 69:361-370.
D. J., J. F. Kitchell, and L. B. Crowder. 1981. Forage Fishes and Their
Salmonid Predators in Lake-Michigan. Transactions of the American Fisheries
T. M., J. A. Estes, D. F. Doak, and A. M. Springer. 2004. Killer appetites:
assessing the role of predators in ecological communities. Ecology 85:3373-3384.
D., R. Froese, and D. Pauly. 2005. On losing and recovering fisheries
and marine science data. Marine Policy 29:69-73.