Synecology of species-rich plant communities on roadside verges in the Netherlands
Schaffers, André P.; Sýkora, Karlè V.
Using a large number of physical and chemical soil measurements, biomass measurements, and other site conditions (e. g. management, shading, exposition), an accurate synecological description is given of 15 semi-natural, species-rich plant communities. The communities studied belong to 11 alliances, comprising 10 associations and 3 trunk communities: Urtico-Aego podietum, Alliario-Chaerophylletum, Valeriano-Filipenduletum, Fritillario-Alopecuretum pratensis, Calthion, Triglochino-Agrostietum stoloniferae, Ranunculo-Alopecuretum geniculati, Arrhenatheretum elatioris, Mesobromion, Phleo-Tortuletum, Spergulo-Corynephoretum, Genisto-Callunetum, and Ericion. The studied syntaxa, covering a wide range of (unfertilized) edaphic conditions, represent a cross-section of the more valuable plant communities occurring along roadsides in the Netherlands. The study thus renders a synecological reference framework, valuable not only for practical applications such as ecological roadside management and (re-)construction but also from a broader scientific interest since all the communities studied occur also outside roadside habitats and several also occur outside the Netherlands. To guide the synecological descriptions, the data were analysed for 'master factors', defined as those environmental variables that best distinguish between the plant communities. Discriminant analysis revealed eight: shading intensity, spring groundwater level, average lowest groundwater level, soil pH, CaCO3 content, the fraction of soil particles < 16 μ-m, the degree of nitrification, and the amount of available K. Of the studied sites, 87 % could be accurately classified using these environmental variables and only 4 % were misclassified as entirely unrelated syntaxa. It was concluded that the obtained master factors permit an adequate ecological description and it is suggested that for practical applications assessment of these eight variables is sufficient.