Effects of fishing on fish assemblages in a coral reef ecosystem: From functional response to potential indicators

• Models coupling spatially explicit data of fishing pressure and reef fish metrics were developed to analyse ecological responses to fishing.
• The description of fish assemblages through a panel of metrics was essentially based on a functional approach rather than a taxonomic approach.
• Accounting for the natural variability of assemblages using multi-scale factors allowed detecting impacts despite moderate fishing contrasts.
• Several metrics describing reef fish resources provided potential indicators of fishing impact for moderate levels of fishing pressure.
• Similar approaches are much needed for the preventive management of resources and ecosystems when significant changes in fisheries are expected.

Many Pacific countries face socio-economic changes, resulting in an increase in fishing pressure on reef fish resources. In coral reef environments, examining the ecological effects of fishing on fish assemblages is generally hampered by the lack of data for both in situ assemblages and fishing activities, and by the difficulty to account for natural variations in assemblages. The responses of fish assemblages to moderate fishing pressure were investigated in two sites of New Caledonia (South Pacific). Habitat structures and fish assemblages were described through in situ multiyear underwater visual surveys to account for spatial and temporal variability. A recent quantification of fishing catches and efforts allowed defining spatial gradients of fishing pressure. Linear models were used to test fishing effects on species richness, density, biomass, length and trophic structure of the whole assemblage, large species, target species, and different ecotrophic guilds. A significant decrease in most metrics was detected when fishing pressure increased. The results obtained for all-species fish metrics were confirmed for target species, large species, carnivores and corallivores trophic guilds and large individuals. This functional modelling approach accounted for the natural variability of local coral reef systems and allowed detecting effects of fishing, although fishing pressure was low to moderate in the surveyed sites. Our results suggest that further modelling development, including control sites, would allow precautionary indicators and strategies to be defined for the monitoring and management of reef fish resources.