Simulating and testing the sensitivity of ecosystem-based indicators to fishing in the southern Benguela ecosystem

The sensitivity of size-based, species-based, and trophodynamic indicators is examined for the fish community of the southern Benguela ecosystem (South Africa) through simulations of different fishing scenarios using the multispecies model OSMOSE. The simulations suggest that it may be erroneous to consider one absolute reference direction of change for any indicator because the direction of change is specific to both the multispecies assemblage and the fishing scenario considered. The analysis of species versus community indicators is helpful for understanding which processes drive the emergent properties of the ecosystem. Informative about the structure and state of the ecosystem, both types of indicators weighted by biomass or by abundance should be used to evaluate ecosystem changes. Indicators characterizing size distribution (e.g., slope of size spectrum) appear to be more helpful in distinguishing the cause of ecosystem changes than mean community indicators because their response is specific to the fishing scenario simulated (i.e., global or hake-targeting fishing). Some indicators do not seem to be sensitive to fishing pressure (slope of the diversity size spectrum) or do not vary consistently with other studies (W statistic)