Long-term microphytoplankton variability patterns using multivariate analyses: ecological and management implications.

Microphytoplankton data along the southern Tunisian coasts have been monitored weekly since 1995. This study used the data collected during the period 2000-2007 to determine the temporal variability patterns and to characterize the geographical structures of microphytoplankton populations. The methodological approach consisted in isolating the temporal variability common to all sampled sites and in analyzing the "residual" site-specific component. Multivariate ordination methods were used to determine the seasonal and interannual variability. The temporal pattern shared by all sites was highlighted and the seasonal cycle of the most frequent phytoplankton communities contrasted the diatoms winter-spring species to the dinoflagellates summer-fall species. Multitable comparison allowed the identification of three areas exhibiting similar phytoplankton population variability. The results, despite the coastal location of sampling sites, indicated that hydrodynamic and geomorphologic properties of the different areas as well as anthropogenic activities play a key role in the structure of the phytoplankton communities.