A compositional analysis approach to phytoplankton composition in coastal Mediterranean wetlands: Influence of salinity and nutrient availability

Mediterranean wetland communities are strongly constrained by hydrological perturbations and the water flow regime. Salinity and nutrient availability have often been considered the most important variables determining changes in the phytoplankton community of coastal wetlands. Ratios between the main environmental variables often have more relevance than the absolute values of each variable; however, in most cases ratios are not suitable for use in multivariate models commonly used by limnologists. The main objective of the present work was to identify the main variables or variable ratios that are the driving forces of the major phytoplankton taxonomic groups in Mediterranean coastal wetlands, using compositional data analysis techniques (CoDa). With this aim, eleven shallow wetlands (6 in Empordà, 5 in Doñana, NE and SW of Spain respectively) were sampled in winter and spring 2007. Two approaches were used: the first one using raw data and the second one using CoDa techniques to transform data. Our results show that differences in hydrological patterns led to three main community assemblages, ranging from communities dominated by typical marine taxa (diatoms and dinoflagellates) when the marine influence was high, to communities dominated by cyanobacteria during confinement and when inorganic nitrogen was scarce. In freshwaters with a high turnover rate, the community was dominated by opportunistic chlorophytes and cryptophytes that need inorganic nitrogen availability. When the raw data and CoDa approaches were compared, the CoDa approach permitted a better ecological interpretation of the phytoplankton community and the main ecological processes. Salinity was the main environmental determinant with both approaches, while the second CoDa RDA axis was related with the balance between the peptidase and phosphatase enzyme activities, confirming the relevance of nutrient retrieval processes in determining phytoplankton composition. We recommend the use of CoDa techniques for analyses of planktonic communities, such as the one presented here, in order to improve the interpretation of both existing and future datasets.