Some statistical tools for the analysis of lagoons and aquatic systems data

In this work we propose a review of researches developed in recent years by statisticians and biologists working on lagoons and coastal areas. We briefly illustrate the use of: (1) spatial statistical tools to fill gaps in the data, (2) linear models (LM’s) and GAM’s to investigate the relationships between food web metrics and environmental variables, (3) LM’s, GAM’s, GAMM’s and linear mixed models to investigate the relationships between ecological multimetric indices and environmental parameters and (4) a model based approach for the classification of the ecological status of lagoons. Points (1) and (2) have been developed in collaboration with the Environmental Biology department in Rome, based on data from an aquatic ecosystem in South Sardinia, in order to address two questions: to what extent the effect of disturbance in the terrestrial landscape spreads through the nearest aquatic ecosystem? Which aquatic community metric is more suitable to indicate the propagation distance of terrestrial disturbance? The other two topics come from the collaboration with Environmental Biology department at university of Salento, working on the classification of the ecological status of Mediterranean and Black Sea lagoons. Here the goal is to understand the behaviour and classification ability of multimetric indices (MAMBI, AMBI, BAT, BITS and ISS). In (4) we try to solve the disagreement often encoutered among alternative lagoon classifications, proposing a joint Bayesian hierarchical model for four multimetric indices, based on multivariate normal mixtures. In this model each mixture component is estimated as a function of covariates of interest and corresponds to an ecological status.