Search

An age-aggregated model that explicitly includes cannibalism of older fish on younger fish is developed. This model is used as the basis for determining the importance of including cannibalism in stock assessment and fishery regulation for the Cape hake stocks off southern Africa, stocks where cannibalism has been shown to be substantial. The cannibalism model is compared to a traditional Schaefer-type model in three ways. First, the two models are fitted to the catch-effort history for the Cape hake resource off northern Namibia. It is shown that the cannibalism parameters cannot be estimated from the catch-and-effort history alone, and must be obtained from other studies. Once the cannibalism parameters are specified, the two approaches provide similar fits and estimates of stock size and potential yield, although the non-cannibalistic approach is generally more optimistic. The second form of comparison involves fitting a cannibalistic model-estimation procedure and a non-cannibalistic model-estimation procedure to simulated data that contains differing levels of cannibalism. In general, both model-estimation procedures similar results, although the non-cannibalistic model-estimation procedure is more likely to result in outlying estimates of MSY when the data are uninformative. Finally, the performance of the two model-estimation procedures when combined with the f_MSY harvesting strategy is evaluated. The performance of the two approaches is similar as long as the data are relatively informative. It is thus argued that management agencies should not assume a priori that including cannibalism or multispecies interactions in assessment and management procedures will necessarily provide for better fisheries regulation. It is recommended that simulation studies similar to those described in this paper be performed prior to the introduction of the extensive data collection schemes required for assessments that include cannibalism or multispecies interactions. Copyright 1994, 1999 Academic Press