Modeling the consequences of antarctic krill harvesting on antarctic fur seals

In terms of the convention governing the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR), management advice for the Antarctic krill (Euphausia superba) fishery should take the needs of the predators of krill into account in order to reduce the risk of deleterious impacts on such predators (e.g., baleen whales and numerous fish, seal, penguin, and flying bird species). A single species yield model is currently being used by the CCAMLR Scientific Committee to recommend an annual krill catch limit, which is expressed as a proportion (γ = 0.116) of a survey biomass estimate. This approach takes the needs of predators into account in only a crude way by assuming that a median krill escapement of 75% of its unexploited biomass would be sufficient to meet the needs of predators. A krill–predator modeling procedure is presented that could be used to directly assess the impact of krill harvesting on krill predator populations and therefore to revise this recommended harvesting level (γ) for Antarctic krill. Application of a deterministic form of this model to an Antarctic fur seal (Arctocephalus gazella) data set from Bird Island, South Georgia, Antarctica, indicates that the level of krill fishing intensity (γ) that would reduce this population to half the equilibrium size in the absence of krill fishing (γₕₐₗf) lies between 0.03 and 0.18, which includes the level recommended by CCAMLR. This large range results primarily from the sensitivity of the model to the maximum growth rate parameter, for which a range of 5–15%/yr is investigated. A plausible range of values for this parameter (5–15%/yr) results in estimated γₕₐₗf values from 0.04 to 0.23. Stochastic calculations (which take account of interannual fluctuations in the abundance of the krill population due to recruitment variability) yield higher estimated γₕₐₗf values than the less realistic deterministic calculations. However, simulation tests indicate that the estimated γₕₐₗf values are biased upward; this modeling approach is therefore likely to yield γₕₐₗf values that would lead to a depletion of the Antarctic fur seal population to more than half its pristine size.