Workshops on Apex Predators in Marine Systems:
Final Report to the National Center for Ecological Analysis and Synthesis

The world's marine and freshwater fisheries give strong evidence of extensive and continued overexploitation. That raises important questions about the role of fishing as a major ecological force that can substantially alter the demographic characteristics of targeted species and their role in a food web or ecosystem context. In an ecological sense, one can view fishery exploitation as a "press" experiment and the changes evoked by strong management actions as a "pulse" experiment. While the fishery catches may be monitored over time and offer the benefit of extensive, long-term data sets, the challenge is to find fishery-independent evidence of the suspected and larger-scale ecological effects.

These workshops were organized by Jim Kitchell as part of his sabbatical appointment at NCEAS during the period of Jan.-June 1997. The following is a brief report on the first workshop held during 30 January-1 February 1997 at the NCEAS facilities in Santa Barbara.

The general goal of this project is to: Evaluate the ecological effects of fishery exploitation on apex predators in marine systems and to deduce how those might be expressed in their food webs.

Results of the First Workshop: 30 Jan. -1 Feb. 1997

All of the first day and part of the second were dedicated to reviews of the main question as pertinent to each of a series of "case study" ecosystems. In all case studies, the reviews focused on the history of the fisheries, the current state of understanding of food webs, and the likely ecological changes effected by previous or intended fishery management actions. Those covered are listed below:

"Central Pacific/Hawaiian Islands Longline Fisheries for Billfishes and Tunas." Chris Boggs, National Marine Fisheries Service, Honolulu HA "Eastern Tropical Pacific Purse Seine Fisheries." Robert Olson, Inter-American Tropical Tuna Commission, La Jolla CA

"Gulf of Mexico Bycatch Issues and Chesapeake Bay Fisheries" James Cowan, Dauphin Island Sea Lab., Mobile AL

"Baltic Sea and North Sea Fisheries and Food Webs" Sture Hansson, Univ. of Stockholm, Sweden

"Gulf of Maine Cod-Lobster Interactions, Recovery of the East Coast Striped Bass Populations, and Georges Banks Fisheries and Food Webs" Xi He, Massachusetts Marine Fisheries Commission, Gloucester MA.

"Trophic Cascades in the Bering Sea" Robert Francis, Univ. of Washington, Seattle WA

Three classes of general tools and approaches were envisioned for future analytical work. Those included:

  1. Bioenergetics models, which were already familiar to the participants,
  2. Demographic/life history models as pertinent to sharks (reviewed by Greg Cailliet, Moss Landing Marine Laboratory, Monterrey CA), and
  3. Ecosystem/food web models (Ecopath and Ecosim)(reviewed by Carl Walters, Univ. of British Columbia, Vancouver, Canada)

A full day of general discussion followed the case history presentations and reviews of models. Each participant also prepared a 2-3 page written summary of their contribution.

Several follow-up activities were identified as having greatest potential and specifically planned for post-workshop effort. Cowan will develop a more contemporary version of the previous Ecopath analyses conducted on Gulf of Mexico shrimp fisheries. His collaboration with Walters will allow development of an Ecosim model for future analyses. Working from Olson's extensive data base on diets of apex predators, Boggs, He and Kitchell will develop an Ecopath/Ecosim model for the pelagic systems adjacent to the Hawaiian Islands. Hansson will work on developing "hindcasting" of food web linkages based on the use of energetics models to track bioaccumulation of contaminants (e.g. PCBs) and/or stable isotopes (N and C). Francis will pursue the prospect of developing an Ecopath/Ecosim approach targeted on the question of carrying capacity for salmon in the N. Pacific. Cailliet will continue to accumulate background on the development of and proposals for Marine Fishery Reserves as the possible basis for future, ecologically-focused work.

Submitted by Jim Kitchell
6 February 1997

Report of the Second Workshop: 23-24 May 1997

The second workshop on Apex Predators in Marine Systems was convened during the period of 23-24 May 1997. As a result of the first workshop, a revised version of the Ecosim model (Ecosim 2) was developed to incorporate two important additions pertinent to the interest in density-dependent effects. First, a delay-difference equation system was implemented to allow for distinction between adult and juvenile stages as part of the recognized need for attention to life history differences among apex predators. Second, density-dependent growth was included as a means for assessing ecological responses not directly represented by changing mortality schedules. These modifications are described in greater detail in a manuscript recently submitted to the Canadian Journal of Fisheries and Aquatic Sciences (Walters, C. J., D. M. Pauly and J. F. Kitchell. Representing density dependent consequences of life history strategies in an ecosystem model: ECOSIM II.).

Initial versions of Ecopath models required to run Ecosim 2 derived from the first workshop and the follow-up provided by its participants. Most of the modeling work was conducted during the period of April through June while Kerim Aydin, a graduate student working with Robert Francis at the Univ. of Washington, was sited at NCEAS. Aydin was jointly supported; the U. Washington Sea Grant Program provided a salary stipend and NCEAS provided travel funds plus a housing allowance.

Three ecosystems were chosen for more detailed analyses: 1. The Gulf of Mexico, 2. The Alaska Gyre, and 3. The Central North Pacific. A preliminary report on those models and their results was presented by Kitchell at the National Sea Grant Week during August of 1997. The background and key results for each are summarized in the following.

Gulf of Mexico

The Gulf of Mexico model was implemented to evaluate two important ecological issues: a. the effects of bycatch in the shrimp trawling industry and their long-term consequences for demersal and pelagic fish populations, and b. the likely scenarios for large-scale change due to implementation of bycatch reduction devices (BRDs) which, like turtle excluder devices (TEDs) are designed to screen juvenile and adult fishes from the catch and, therefore, reduce their mortality rates due to bycatch. Ecosim model simulations produced the expected result that demersal fish populations would increase over about a 20-year period after BRDs were employed in the fishery. An unexpected result was that shrimp and crab fisheries did not decline strongly due to increase in one of their predators (demersal fishes) which we interpret as due to the high productivity of these short-lived crustaceans. A second, unanticipated result was that shark populations increased due, primarily, to the enhanced prey resource represented by increasing abundances of demersal fishes.

Follow-up Activities
Kitchell was invited to present highlights of this work at an October '97 workshop on Ecology and Conservation Biology of Large Pelagic Fishes which was jointly sponsored by The Billfish Foundation and The Marine Conservation Biology Institute. As a result of that presentation, members of the Gulf Coast Marine Fisheries Council are currently developing plans for follow-up modeling workshop(s) sponsored through their auspices. The Ecopath used in this analysis was "on loan" from its developer, Dr. Joan Browder, NMFS, to Jim Cowan as a basis for the preliminary Gulf of Mexico Ecosim model. Papers describing that Ecopath are forthcoming, but those describing the Ecosim results can only be used for heuristic purposes until the original documentation is in print. The ecological issues embodied in this model and the consequences of regulations requiring BRDs are contentious and of substantial economic importance. This version of the model and its applications will develop slowly and cautiously.

Alaska Gyre

Aydin developed the Ecopath version employed in characterizing the Alaska Gyre ecosystem which is the feeding grounds of most of our N. Pacific salmon stocks. Simulations were conducted to evaluate the relative importance of salmon stocking policy and the large-scale changes thought due to oceanic regime shifts which is thought to have increased primary production in the higher latitudes of the N. Pacific. A simulated "regime shift" increase primary production by approximately 50% which was transferred up to all levels of the food web and increased salmon stocks as expected. When the regime shift was concluded after a 20-year period, an unexpected result developed: salmon dropped to much lower biomass levels than those that preceded the regime shift. The model results ascribed this decline to predatory inertia which owed to two decades of gradually increased biomass of slowly growing salmon predators such as salmon sharks. Thus, when the environment changed to lower primary production levels, the increased mass of predators imposed a greater mortality on salmon and evoked the consequent reduction in salmon biomass.

Follow-up Activities
Aydin will fully develop and document this version of Ecosim as part of his Ph.D. thesis at U. Washington. The 1997 El Nino will probably stimulate additional work with the model as Aydin and Francis rise to the challenge of forecasting immediate and long-term consequences of El Nino effects on ocean carrying capacity.

North Central Pacific

The primary question developed in this model focused on the effects of the large and growing fleet of longline fishers operating in the central Pacific. Bycatch of sharks, turtles and birds (e.g., albatross) are of growing concern to conservationists. The increasing exploitation of apex predators are of concern to ecologists and fisheries managers. Boggs and Xi developed the initial Ecopath version for this model. Kitchell and Aydin subsequently worked on the Ecosim 2 version that derived from this.

Food webs of the Central North Pacific pelagic ecosystem are characterized by a guild of apex predators including several species of billfishes, tunas and sharks. Longline fisheries impose size and species-selective exploitation on those predators which may alter community composition and cascade to lower trophic levels. We used a version of Ecosim 2 that was developed to evaluate the ecosystem effects of fishing at different and sustained rates over a 50-year period. Scenarios were designed to reveal the role of selective and non-selective fishing strategies in altering: 1. relative abundance of individual apex predator species, 2. responses of key prey species to changes in exploitation of their predators, and 3. effects of fishing on the intensity of intra-guild cannibalism.

Life history characteristics are a first and important predictor of responses as exploitation rates increase. Species characterized by slow growth rates, delayed maturation and low fecundity (e.g. sharks) decline first and recover slowly when fishing rates are reduced. Those with rapid growth, early maturation and high fecundity (e.g., skipjack and yellowfin tunas, mahimahi) increase as larger predators (e.g., billfishes and sharks) are removed by fishing. Reduced predation allows mesopelagic prey and flying fishes to increase, but compensatory increases in survivorship among the juvenile life stages of all tuna and billfish species tend to minimize responses by key prey species such as epipelagic fishes and squids. Under conditions of intense exploitation (2X current levels), equilibrium adult biomass of many species decline and remain low. However, adult mahimahi biomass recovers quickly and eventually increases to severalfold the initial conditions. Biomass of juvenile yellowfin, bigeye and albacore tunas plus that of juvenile billfishes increase under conditions of intense exploitation while those of juvenile skipjack and mahimahi decrease.

When exploitation is removed, biomass of all juveniles declines rapidly as does biomass of adult mahimahi. Adult biomass of all other species increases. Sharks increase gradually, but the largest and most rapid relative increases occur in yellowfin tuna, swordfish and other billfishes. Collectively, the simulation results demonstrate expected effects on adult stocks, but counter-intuitive responses for the juveniles because exploitation diminishes intra-guild cannibalism and increases juvenile survivorship in many species.

Follow-up Activities
Kitchell and Boggs are pursuing prospects for continued funding of analysis of the Central North Pacific ecosystem through two efforts. Kitchell traveled to Hawaii in April where they developed the basis for a proposal "An Ecological Analysis of Apex Predators in Pelagic Systems" which is now pending with the Biological Oceanography Program at NSF. Kitchell will travel to Hawaii in early 1998 to work with Boggs in writing a manuscript describing this Ecosim application and its interpretation.

Bob Olson has initiated efforts to develop a version of Ecosim pertinent to the Eastern Tropical Pacific. That effort will continue through the auspices of the Inter-American Tropical Tuna Commission.


In summary, the workshop on Apex Predators in Marine Systems were intended to create a focus on approaches that might be employed in evaluating otherwise disparate views of the role of fishery exploitation in an ecosystem context. Several tangible results derived.

  1. Kitchell and Francis now sit as members of the National Marine Fisheries Service Panel on Ecosystem Principles. The Panel's report is in preparation and will be presented by October, 1998, as a source of guidance to development of NMFS management strategies.
  2. The Ecosim model has been developed in a variety of applications. The software is now available from the Fisheries Centre, Univ. of British Columbia, and will be made available with the new version of Ecopath offered through the International Commission for Living Aquatic Resources based in Manilla, Phillipines.
  3. Highlights of the Ecosim applications developed for the Gulf of Mexico, Alaska Gyre and Central North Pacific ecosystems were presented at a national meeting (National Sea Grant Week) and in an international forum (Workshop on Ecology and Conservation Biology of Large Pelagic Fishes).
  4. Additional development of Ecosim models includes one recently derived for Lake Superior when Aydin visited U. Wisconsin during September of this year. That model will be the focus of a workshop with the Lake Superior Fisheries Technical Committee to be held in Madison during 10-11 November 1997.
  5. A proposal (An Ecological Analysis of Apex Predators in Pelagic Systems) has been submitted to the NSF Biological Oceanography Program.
  6. A manuscript describing the Ecosim 2 model has been submitted for publication.
  7. The collegial network established through these workshops has become a major source of ongoing, creative and productive collaborations.

Submitted 3 November 1997
James F. Kitchell