Abstract
Coastal marine ecosystems are of utmost importance to Humans due to their economic, social and cultural values. Yet, they are heavily impacted by our actions which reorganize species assemblages by altering their species richness, their composition and the abundances of species present. Evaluating how these reorganizations affect ecosystem functioning can be achieved by measuring the functional diversity of these assemblages. Most of the coastal marine ecosystem diversity is made by bony fishes which play a crucial role in their functioning. Therefore, this thesis studied how human activity impacts the fish's functional diversity of coastal marine ecosystems. To better understand the functional role of fishes, I also developed new tools for studying the temporal variability of biodiversity.I coordinated the development of a R package called mFD. It includes 18 functions to compute and plot all functional diversity indices in a reproducible and reliable manner. Then, I proposed a new methodology for observing marine diversity and its temporal variability. It uses a network of synchronized cameras capable of filming a standardized area for over 12 hours. Lastly, I developed new indices to account for the temporal variability of species abundance to better understand their functional role. I developed the SmaxN metric which estimates the abundance of each species in a habitat using synchronized cameras. I then conceptualized and tested three metrics characterizing species temporal rarity by quantifying the amount, variability and regularity of its abundance over time.I used these methodologies to answer several questions in ecology. First, I focused on biological invasions and their impact on ecosystem functioning. In a Mediterranean Sea area where the two native and the two invasive herbivorous species coexist, I demonstrated that the herbivory process was dominated by one invasive and one native species, and that the presence of the invasive species does not impact the herbivory activity of the native species. While this study highlighted a high variability of the herbivory process between and during days, it also validated the hypothesis of a more intense grazing activity in the afternoon. Second, I focused on current marine conservation of marine biodiversity by measuring the effect of the levels of protection in Marine Protected Areas on ecosystem functioning. By studying Mayotte coral reefs (Indian Ocean), I showed that the corallivory and herbivory processes can be more important at lower protection levels, and that feeding activities studied are highly variable on an inter and intraday scale. I finally studied the temporal accumulation of taxonomic, functional and phylogenetic diversities over the sampling days and demonstrated the importance of considering the inter and intraday variability which can bias our evaluation of conservation measures due to the presence of temporally rare species.