Análise da variação na composição das comunidades marinhas dos costões rochosos da Baía da Ilha Grande e investigação dos potenciais agentes promotores da biodiversidade

Abstract

What makes communities different within a metacommunity? Understanding the drivers of biodiversity variation remains a challenge. Moreover, marine systems are usually understudied when compared to their terrestrial counterparts, which, being more accessible, have been investigated for much longer. The question around the mechanisms underpinning community variation within a system, called beta diversity (β), has been historically linked to the development of niche theory. The communities were considered species associations controlled mainly by the environmental conditions and biotic interactions found at the site. Therefore, this set of conditions was thought to restrict which species were capable of getting established and thrive in which regions in a way that minimise niche overlapping. Later, the influence of other local factors not related to niche, such as local random extinction, as well as processes not connected to the local communities but rather dependent on regional- scale constrictions, like immigration, were recognise as further drivers of diversity. It was our goal to better understand the marine rocky shore community structure from Ilha Grande Bay (BIG in the Portuguese acronym), Brazil, and relate the observed patterns to their potential causes, whether linked to niche or not. The presence/absence data from 773 species of six taxonomic groups, benthic macroalgae (110 species), Cnidaria (26), Echinodermata (27), Crustacea (61), Mollusca (374) and reef fish (175) previously collected in the bay, was used to address those questions. First, it was shown that marine communities in BIG had a high variation rate between sites (β diversity) and such variation was mainly due to species turnover in the spatial gradient. Turnover patterns are usually related to the submission of the species distributions to environmental gradients or stochastic processes. Thus, we assessed how change in the environmental condition acted upon these communities variation using logistic regression in a Generalised Linear Models (GLMs) approach. It was demonstrated that variation in the hard bottom benthic species and reef fish communities could be partially explained by depth differences across the bay, making these communities depict a big scale east-west gradient. Meanwhile, variation in soft bottom organisms communities could not be explained by any given environmental variable. For all six taxonomic groups, most of variation also remained unexplained. After this, using algorithms capable of generating matrix randomisations it was possible to verify that competition was not one of the key factors driving community composition, although it was observed for some specific situations. Furthermore, most of the sites sampled did not differ from random expectations in regard of number of shared species between sites, meaning they were no more different than expected by chance. In summary, it was possible to find that most of variation among these marine rocky shore communities at BIG could not be explained by any measured deterministic driver and apparently change in community composition is random at this scale. An alternative to the usage of the traditional canonical analysis for community ecology studies was tested and was found to perform better