Link to NMFS Homepage Link to NOAA Homepage Keyboard Navigation Alaska Fisheries Science Center Home banner


Resource Ecology and Ecosystems Modeling Program

Laboratory analysis was performed on 3,198 groundfish stomachs from the eastern Bering Sea and 143 from the Gulf of Alaska. A total of 1,349 stomachs were collected from the Aleutian Islands region and 10,447 from the eastern Bering Sea. Five observers returned 474 stomach samples from the Bering Sea during the quarter.

Sensitivity analysis of the multispecies virtual population analysis model parameterized for a system of trophically-linked species from the eastern Bering Sea.
A sensitivity analysis of a multispecies virtual population analysis (MSVPA) model parameterized for the Bering Sea was carried out by Jesus Jurado-Molina.  Two methods, a fractional factorial design of resolution IV and an individual perturbation parameter method were used for the analysis. Input parameters varied were residual natural mortality (M1), predator ration, and terminal fishing mortality of fish species in the MSVPA.  Four variables related to the predation mortality (M2) and population size of age-1 walleye pollock and Pacific cod were chosen as response variables.  We considered a response variable being sensitive to a predictor variable when a 10% perturbation produced a percent change bigger than 10%. Most response variables were not sensitive to input parameter perturbations, with only one individual parameter perturbation resulting in a change greater than 10%.  This was the perturbation of Pacific cod residual natural mortality that resulted in a 15% change in the number of age-1 cod.  Results suggested that the response variables of walleye pollock were only slightly sensitive to changes in some input parameters, including population input parameters of walleye pollock and ration and population number of arrowtooth flounder, with responses ranging from about 3% to 6%. Pacific cod response variables were sensitive only to variables related to Pacific cod with responses ranging from about 6% to 9% for parameter perturbations of cod terminal fishing mortality, ration and residual natural mortality. Other parameter perturbations produced changes in pollock and cod response variables of less than 3%.  

Results from the individual perturbation parameter analysis also showed that the changes produced by large positive perturbations in the other food parameter were small. The results obtained in this work reinforce the overall conclusion of the robustness of the MSVPA found in previous works. They are also an important step in the validation of the MSVPA and MSFOR models and help in the identification of the parameters requiring further refinement, including the improvement of estimates of annual ration of predators and residual mortality. This model validation is necessary to incorporate multispecies models in ecosystem-based fishery management advice.  However, it is also necessary to develop statistical multispecies models able to assess the uncertainty of parameters producing the largest effects in the MSVPA model.


Uncertainty in Trophic Cascade Models
In conjunction with NMML researchers, Kerim Aydin prepared an analysis of the uncertainty inherent in attempts to manage and control trophic cascades in large marine ecosystems.  Specifically, a ‘general’ and extremely simplified trophic model was built in which marine mammals (pinnipeds) competed with a fishery for a groundfish (gadid) species.  Other species in the model included phytoplankton, zooplankton, a generic forage fish, and a fish predator, which was assumed to have no developable commercial value (Fig.2).

The model was calibrated so that, at the start of the modeled time period, the system was in equilibrium, and the fishing yield on the groundfish species was at maximum sustainable yield (MSY).  The model was manipulated by reducing the pinniped population by 50% and allowing the model to reach a new equilibrium over 50 years.  The resulting year 50 biomass levels and fisheries catch (assuming constant fishing effort) are shown in Figure 3.  As can be seen, the reduction in pinnipeds did indeed increase groundfish biomass and fisheries yield by approximately 10% (bars in Fig. 3).  However, a Monte Carlo simulation of the error ranges for these estimates (assuming very low +/-30% error for input diets, +/-10% error for initial biomass and production levels) resulted in the broad 95% confidence intervals shown by the error bars in Figure 3.  As can be seen, groundfish biomass and catch were likely to go down or up, and estimates of juvenile groundfish biomass range from explosively upward to crashing.

This result is explained by the fact that the removal of a slow component of the ecosystem (marine mammals) helps its competitors (faster-growing predatory fish of low commercial value) as often as it helps its prey (groundfish).  This is true with the smallest realistic levels of input error (equilibrium assumptions, unrealistically small range of error in diets).  In other words, trophic cascades are not possible to control across a large marine ecosystem (stock-level) management scale even with extremely precise data; rather, trophic manipulations must be directed and localized in time and place, to avoid causing undesirable and unexpected changes in the system.  New models are being prepared to include manipulations involving lower trophic level marine mammals (baleen whales) and are being extended to more detailed models (for example, the eastern Bering Sea model as modeled with 40+ functional groups).

By Pat Livingston.


Economic and Social Sciences Program

Bering Sea/Aleutian Islands Crab Rationalization
The mandatory reporting requirements for economic data were further developed through an industry and interagency effort, and the results of that effort were identified in a report prepared for the NPFMC.

Fishing Capacity
A report, “Quantitative Estimates of Fishing Capacity, Capacity Utilization, and Fishery Utilization for Alaskan Commercial Fisheries, 2001” was completed as part of an agency-wide effort to estimate excess fishing capacity for all federally managed commercial fisheries. A significant difference between capacity and actual catch may signal the need for implementing measures to diminish or eliminate the incentives for, and presence of, excess capacity.

The capacity measures computed in the report were constructed using data on catch (in metric tons), participation (in weeks), and vessel characteristics of catcher vessels and catcher- processors that operated in federally managed Alaskan commercial fisheries for 1990 to 2001.  Those data were used to estimate fishing capacity, excess capacity, and capacity utilization in 2001 by fleet and species or species group.  Fishery utilization was estimated for each fleet and estimates of the number of vessels and mean number of fishing weeks by fleet were reported for 1990-2001.

There are wide ranges of fishing activities, vessel sizes, targeting strategies, and gear configurations in the various federally managed Alaskan fisheries.  Generally speaking, however, groups can be established that are likely to share similar technological, economic, and regulatory constraints (TACs, closures, seasonal delineation).  In an attempt to establish such groups, vessel characteristics, fishery participation, and processing data (for catcher-processors) were examined.  As a result, 12 catcher vessel fleets and 10 catcher-processor fleets were formed.  Each of these fleets is comprised of similarly equipped, similarly sized vessels that engage in a common set of fisheries (and, in the case of catcher-processors, produce a similar set of finished products).  Such a grouping allows us to present the capacity estimates on a fleet-by-fleet basis, which more clearly elucidates the sources of fishing capacity.

The estimates indicate that current capacity, in terms of total catch of all species, exceeds actual catch by nearly 40%.   However, species-specific excess-capacity estimates range widely, from 8% to over 300%.

Subsistence Hunting and Fishing Practices
Jennifer Sepez presented a paper at the Ninth International Conference on Hunting and Gathering Societies in September in Edinburgh, Scotland.  The paper “If Middens Could Talk: Comparing Ancient, Historic, and Modern Makah Subsistence Foraging Patterns” combined archaeological data with data from early ethnography and contemporary harvest surveys to examine consistency and change in Makah subsistence hunting and fishing practices between 1500 and today.

The data indicate a significant shift in contribution of different resource groups to the animal protein diet between 1500 and today, with harvest of marine mammals  dropping tremendously (from 92% to less than 1%), and the contemporary diet consisting primarily of fish (50%), shellfish (11%), land mammals (15%), and store-bought meats (24%).

A high diversity of species used by tribal members prior to Euroamerican colonization are still in use today, from halibut and salmon to harbor seals and sea urchins.  Several species no longer used, such as wolves and fur seals, can be explained by ecological factors, such as post-colonial extirpation. Other resources no longer used, such as many small birds and small shellfish, represent a general contraction of the subsistence diet breadth following the introduction of commercial foods.  As predicted by optimal foraging theory, the resources most likely to be eliminated from the diet are those that rank low in terms of post-encounter caloric return.

Tribal members made use of nearly all available resources in ancient times; additions to the tribe’s subsistence base in modern times were due primarily to the introduction of exotic species such as the Pacific oyster, and local population growth of other species, such as the California sea lion. Road building and habitat changes in the forests increased access to land-based resources, such as deer and elk. Land-based resources in general (terrestrial mammals and commercial meats) increased from less than 1% of consumed animal protein prior to 1500 to close to 40% today.  However, with over 60% of animal protein still stemming from marine resources, Makah tribal members remain oriented, both nutritionally and culturally, toward the ocean environment.

Field Research

Social science field research was conducted during July and August in two Alaskan fishing communities by a small team of NMFS social scientists and research assistants.  The group of three spent several weeks talking to community members in Dutch Harbor/Unalaska in the Aleutian Islands, and in Chignik Bay on the Alaska Peninsula.  The team also visited nearby Chignik Lake and Chignik Lagoon, which together with Chignik Bay form a closely-knit complex of villages.  The research locations were chosen as examples of a large fishing community and a small fishing community within the area.  Additionally, both communities have small boat Pacific cod fishers that were differentially affected by Steller sea lion regulations.

More than 80 interviews were conducted with community members including such diverse representatives as cannery workers, vessel captains, community activists, and elected officials.  About 65% of interviews were tape-recorded, which, once transcribed, will allow for textual analysis of data using ethnographic software.  These data is being analyzed.  Reports on the research will include a profile of each community, which will become part of a larger NMFS social science project to develop profiles and data sets on a wide variety of North Pacific fishing communities.

By Joe Terry and Jennifer Sepez.

quarterly July-Sept 2002 sidebar Contents


Auke Bay Lab

National Marine
Mammal Lab

RACE Division

REFM Division

Quarterly Index

Quarterly Home