Dynamics of phytoplankton assemblages: analysis of time series from Tyrrhenian coastal waters

Data from Tyrrhenian coastal waters have been analysed and processed, by adopting and implementing statistical procedures both of standard and innovative kind, in order to study the phytoplankton assemblages and their structure, looking for similar trends, recurrent patterns in abundances, species composition and relationships with environmental variables. Data utilized have been extracted from the Sidimar database, which collects complete information produced in the frame of the coastal marine monitoring program coordinated by the Italian Ministry of Environment for the period 2001 - 2009. The Sidimar database represents an excellent opportunity to use long time series of phytoplankton data with the related hydrological and environmental variables. The first phase of my activity was to analyze the entire Sidimar data set, paying particular attention to nomenclature checking, in order to resume the data availability and to fill in a preliminary list of the phytoplankton taxa. A preliminary product of the work done was a Reference List of the phytoplankton taxa, that may be found along the Italian coastal areas In the second phase, my attention has been focused on the marine-coastal areas of the Campania region. The choice fell on the coastal area of the Campania region, since in this area two sea zones paradoxically coexists, the first investigated from long time (Gulf of Naples), the second (Gulf of Salerno) rather poorly represented in the literature, with scarcity of specific studies relating to phytoplankton communities. The work implied the revision of species lists for the different sites, in order to operate an aggregation to higher taxonomic levels, finalized to reduce the number of taxa. Two different levels of aggregation per sampling station were obtained: high definition (HD) > 120 taxa; low definition (LD) < 60 taxa. In such a way I have got data arrays of plenty adequate size for their use with multivariate statistical programs. At the same time, however, maximum possible information on the composition of phytoplankton populations in their spatial and seasonal variations, have been retained, as proved by statistical comparison between the related HD and LD dissimilarity matrices. The phytoplankton time series so revised were then analyzed in order to reveal seasonal patterns. Principal Coordinates Analysis based on Bray-Curtis distances and Jaccard dissimilarities was applied as exploratory techniques of single station data, in order to define and test periodical patterns, major seasonal variations and possible discontinuities due to apparent taxonomical inconsistencies. Subsequently, all station data sets were considered while maintaining separation between those of the Gulf of Salerno from the other located in the Gulf of Naples. I proceeded to the identification of homogeneous groups of data using as ordination techniques the Cluster Analysis, applied to the Bray-Curtis and Jaccard dissimilarity matrices. Different clusters obtained have to be referred to both "spatial" and "temporal" properties of phytoplankton data, allowing to identify evidence seasonal patterns and to differentiate among sampling locations. The overall results have been then described with appropriate graphics showing the percentage breakdown of months and % occurrence of single sampling stations within each cluster. Finally the Indicator Species Analysis have been applied to the Gulf of Salerno and Gulf of Naples datasets, taking into account the subdivision into homogeneous groups provided by the Cluster Analysis. This innovative analysis has therefore provided evidence of strength and statistical significance of the relationship between species occurrence/abundance, and groups of sites. The main outcome of this work can be therefore represented by the reference lists of indicator species, with their related weight, that characterize the study area from the point of view of seasonality and site location.