Fish harvesting in a variable, productive environment — searching for rules or searching for exceptions?

Abstract

Most fish resources in the world are managed in terms of single species models, but there is a growing awareness of their shortcomings. Neither multispecies fisheries models nor ecosystem models have yet proved more useful than single-species models to help manage fisheries, but they are considered to have the potential to provide better advice in both the long- and short-term. This paper addresses the impasse between the single-species, statistical approach and the multispecies, ecological approach. Will understanding ever lead to better management models? A comparison of the spatial dimensions, primary productivity, zooplankton biomass and fishery yields in different productive areas of the world indicates mat the Benguela, the Canary and the California currents are much less "efficient" in terms of fish production than the Humboldt Current or shallow shelf systems in temperate regions such as the North Sea, Georges Bank or the Bering Sea. Reasons for the discrepancies are advanced and the forcing mechanisms and limitations of fish production in the Benguela are outlined in terms of food limitation, critical gateways, mismatch of phytoplankton and crustacean zooplankton and advective losses. The high apparent variability of key processes at small scales of space and time, equivalent to weather fluctuations, to some extent mask longer-term variations. Some suggestions on monitoring, based on good understanding of food production and consumption processes related to pelagic fish, are proposed to modify the assumption of average recruitment in exceptional years, and to provide data on the effects of longer-term climatic changes on productivity in South African coastal waters.