Due to the cost and time required for species identification, the taxonomic sufficiency approach has been developed in order to detect community response to a disturbance, using high taxonomic level without great loss of information. This concept has been widely applied to pollution monitoring studies but rarely to other forms of perturbations such as anthropogenic exploitation of marine resources. We applied this method both to soft-bottom (seagrasses) and hard-bottom (coralline) tropical invertebrate communities in New Caledonia, South Pacific. The objective was to test whether intermediate or high taxonomic levels (genus, family, class or phylum instead of species) are good descriptors of community patterns and changes in assemblages related to harvesting, by comparing harvested to non-harvested areas for the 2 habitats. We pooled species data into coarser taxonomic categories (from genus to phylum) and showed that matrices at different taxonomic resolutions were highly correlated, particularly for genus and family level for both habitats. Differences between harvested and non-harvested locations appeared to be clearly habitat-dependent; for soft habitats, genus and family resolution allowed the detection of changes between exploited and protected assemblages, while for hard habitats, the separation between harvested and non-harvested areas was less clear at high taxonomic level and required species-level identifications. These results suggest that the taxonomic sufficiency approach could be carefully applied to poorly known environments. Family level is a good descriptor of community composition for tropical reef invertebrates. Detecting changes due to anthropogenic exploitation requires different taxonomic resolutions depending on the considered habitat.
KEY WORDS: Taxonomic resolution · Data transformation · Tropical reef invertebrates · Harvesting effect · Marine protected areasResale or republication not permitted without written consent of the publisher Mar Ecol Prog Ser 406: 211-222, 2010 patterns (Ellis 1985), in particular for disturbancerelated changes in benthic communities (Somerfield & Clarke 1995). This method uses coarser taxonomic levels when performing community comparisons, instead of species level. Several papers have addressed the validity of TS in marine ecosystems (Warwick 1988, Ferraro & Cole 1990, 1992, Gray et al. 1990, Warwick & Clarke 1993, Somerfield & Clarke 1995, Karakassis & Hatziyanni 2000. As well as a reduced need for taxonomic expertise, expected benefits include the removal of ecological redundancy attributed to the use of multiple species, when loss of information has no significant effect (Gray et al. 1988, Warwick 1988, 1993. Despite its increasing popularity on impact assessment, TS is still controversial (May 1990, Maurer 2000, in particular in the context of a global decline of taxonomic resources and species knowledge (Scheltema 1996, Boero 2001.Most studies using TS were in marine benthic habitats, including temperate soft-bottom benthic communities suc...