| Resumo: | Benthic habitats are an important component of the marine realm, being this evident by the key role of their associated macrofauna on the marine food webs. Their characterization and mapping assume special relevance in marine ecosystem studies. The main aims of this work were to combine traditional sampling methods with an acoustic remote sensing technique and developing novel modelling approaches, to detail and characterize the benthic habitats and associated macrofauna in the section of the Portuguese continental shelf north of Nazaré canyon. Covering an area of approximately 7000Km2, baseline sediment data and depth were obtained in 226 grab samples, 169 of which were used to study the macrofauna. Also, acoustic transects were run for more than 2500Km. The point sediment data were exploited in two ways: i) using spatial interpolation with Empirical Bayesian Kriging, to produce maps of the sediment descriptors fines%, sand%, gravel% and kurtosis, and of the sediment types classified according to the MeshAtlantic-Folk system; ii) using univariate and multivariate statistical analysis, to relate the sediment data with the spatial distribution of the macrofauna. The acoustic data were collected using the acoustic ground discrimination system QTC VIEW series V, to be used as a surrogate for the environmental data in distribution modelling studies of the benthic macrofauna. Unfortunately, it was not possible to exploit the large amount of acoustic data obtained as it demonstrated to be contaminated by depth, invalidating its use to discriminate sediment types. Nevertheless, the bathymetry data collected with this system allowed to obtain an accurate bathymetric map of the study area, further used as part of the environmental variables layers. Using a wide range of univariate and multivariate data analysis methods, the spatial distribution of macrofauna indices and communities were studied. Distribution models relating the abundance, species richness (or alpha diversity) and Shannon-Wiener diversity with sedimentary parameters and depth were built, using diverse types of regression models. For each biologic index, the predictions from the most accurate model were compared with the predictions of the direct spatial interpolation of the biological data using geostatistical methods. Seven macrofauna communities were identified in the study area, and communities distribution models (CDM) were built, based on binomial regression models, studying the relation of the presence/absence of each community with respect to the environmental variables. Combining the CDM and the maps of the environmental variables, a map representing the distribution of the most probable benthic macrofauna communities was produced. This overall map was used to classify and map the distribution of level 5 EUNIS habitat types and to evaluate their environmental status in the scope of the Marine Strategy Framework Directive, by producing a Marine Biological Valuation map of such EUNIS habitats. The benthic maps produced in this work, either the environmental or the biological, are valuable tools suitable for a range of purposes in the context of the marine ecosystem management, such as monitoring, non-indigenous species control, spatial planning or climate change studies.
|
|---|