• closure and fishing effects;
  • coral reefs;
  • ecosystem management;
  • experimental fishing;
  • fisheries-dependent sampling;
  • Indian Ocean

Abstract  Catch composition, relative abundance and diversity of fish catches in open access and three old fisheries closures were compared and contrasted with previous ecological studies. There was less variation in catch community composition among the fishing grounds than the closures, suggesting that fishing has homogenised catch composition. The trap survey found that some parrotfish [Leptoscarus vaigiensis (Quoy &Gaimard), Calotomus carolinus (Valenciennes) and Scarus ghobban Forsskål] were relatively more common and that some important predators of macro-invertebrates [Balistapus undulatus (Mungo Park) and Cheilinus chlorourus (Bloch)] were less common in the fishing grounds than closures. Unexpectedly, and in contrast to visual census results, cumulative number of species in catch surveys was higher in open access than closures sites. This may result from fishers covering more area and habitat or a reduction in the catch of competitively subordinate and rare species by aggressive, early-caught fish that can dominate bait. Comparisons of ecological visual census surveys and fisheries-dependent methods indicated that small differences in catch composition can reflect larger ecological differences and that baiting methods can underestimate biodiversity. Ecological impacts of fishing and large-scale changes in marine ecosystems must be considerable given the many fisheries-dependent assessments report modest changes.