Large-scale ordination and gradient analysis of salt-marsh communities in the Netherlands in the light of the Dutch National Vegetation Classification

In the development of the Dutch National Vegetation Classification, little emphasis was placed on ordination or gradient analysis. The current study provides a national-scale ordination analysis of salt-marsh communities in the Netherlands and presents it in the context of the National Classification. A series of 4 DCA ordinations derived from synoptic tables of 21 salt-marsh communities, each varying in their emphasis on dominant species, showed that the importance of dominance vs. overall assemblages of species in classifying communities varies with the communities in question: for species-poor communities such as the Spartinetea, Thero-Salicornietea, and portions of the Asteretea tripollii (Puccinellion maritimae and Puccinellio-Spergularion salinae), dominant species were critical in distinguishing communities while for more diverse communities (the Armerion maritimae and the Saginetea maritima of the Astereteae tripollii), total floristic composition was important and over-emphasis on dominant species obscured differences between several communities. Ordination scores were strongly correlated with mean Ellenberg indicator values for salinity and pH, but only weakly with moisture values. In this respect, Ellenberg values do not appear to reflect the critical conditions of duration ad timing of flooding in Dutch salt-marsh vegetation. Ordinations based on both synoptic tables and on individual releves were carried out to study the validity of previously-reported classifications of several related associations. The analysis resolved the question of whether the Suaedetum maritimae is best placed in the Thero-Salicornictea or in the Cakiletea maritimae; it was clearly associated with the Thero-Salicornietea. As for the question of whether certain related associations better belong within the Lolio-Potentillion or within the saltmarsh alliance Armerion maritimae, both options are possible from the point of view of this study. The approach used here of ordinating synoptic tables to locate possible overlap between syntaxa and then using ordinations of individual releves within these to examine the problem areas further provides a useful method for verifying and revising classifications at the national level.