A formal classification of the Lygeum spartum vegetation of the Mediterranean Region

Abstract

AbstractAimsWe examined all available literature and some unpublished data on the We examined all available literature and some unpublished data on the grasslands dominated by Lygeum spartum from Southern Europe and North Africa to produce a formalised classification of this vegetation and to identify the main factors determining its plant species composition.LocationMediterranean Basin and Iberian Peninsula.MethodsWe used a dataset of 728 relevés, which were resampled to reduce unbalanced sampling effort, resulting in a dataset of 568 relevés and 846 taxa. We classified the plots by TWINSPAN, interpreted the resulting pools, and used them to develop formal definitions of phytosociological alliances characterised by L. spartum vegetation. The definitions were included in an expert system to assist automatic vegetation classification. We related the alliances to climatic factors and described their biogeographical features and ecological preferences. The floristic relationships between these alliances were analysed and visualised using distance‐based redundancy analysis.ResultsWe defined eleven alliances of L. spartum vegetation, including the newly described Launaeo laniferae–Lygeion sparti from SW Morocco and the Noaeo mucronatae–Lygeion sparti from the Algerian highlands and NE Morocco. Biogeographical, climatic, and edaphic factors were revealed as putatively driving the differentiation between the alliances. Vegetations of clayey slopes and inland salt basins displayed higher variability in comparison with those of coastal salt marshes.Main conclusionsA comprehensive formal classification, accompanied by an expert system, of the grasslands from Southern Europe and North Africa dominated by Lygeum spartum vegetation was formulated. Eleven phytosociological alliances were recognised, whose plant species composition is influenced by biogeographic, climatic, and edaphic drivers. The expert system, containing formal definitions of the phytosociological alliances, will assist in identifying the syntaxonomic position of new datasets.