RESUMO
The effects of using reduced sampling resolutions to study macroalgal vegetation patterns have not been studied sufficiently. Here, we test the influence of taxonomic resolution level, removal of occasional species, aggregation of species abundances into functional groups and data transformation in the detection of a long-term recovery process by phytobenthic intertidal assemblages. Results indicate that the aggregation of species data into the genus level has very little influence. Likewise, almost any significant information is lost when occasional algae are removed. Analyses at the level of families and orders still clearly detect differences between highly degraded and reference vegetation. By contrast, analyses based on class and functional group abundances capture quite different information. The effect of transformation is similar at the different taxonomic levels. Most surrogate measures properly reflect changes in diversity. It is concluded that genus level is the most appropriate surrogate approach for detecting the recovery process.