• Grassland vegetation;
  • Habitat loss;
  • Land-use;
  • Landscape structure;
  • Phylogenetic signal;
  • Plant traits;
  • RLQ analysis;
  • Three-table ordination



What is the role of landscape structure and dynamics, compared with climatic and geographic factors, in determining species frequencies of grassland plant specialists under habitat loss? Do species traits mediate the relationship between plant community composition and environmental variables?


The Mediterranean mountain grasslands of southern Catalonia (NE Iberian Peninsula), over an area of 100 × 20 km.


Using redundancy analysis (RDA), we explored the association between frequency of broad plant specialists and both present and past habitat patterns in the landscape (i.e. habitat amount and reduction over the period 1956–2003), after accounting for the effect of geographical location and climate in 29 grassland patches. Then, we constructed a database of biological and ecological plant traits potentially related to population persistence, in order to assess the role of these traits in explaining the found association between species composition and environmental variables. We used a single, three-table ordination analysis (RLQ) of the species frequencies, environmental variables and species traits to relate species traits to environmental variables, after allowing for phylogenetic dependence of traits.


The main environmental gradient explaining species frequencies was climatic and geographic. Habitat amount in the current landscape significantly affected species frequencies, while habitat amount in the past landscape did not. A weak but significant association of species traits with environmental variables was detected. Taking into account the phylogenetic signal in plant traits did not change the results.


Plant species in Mediterranean grasslands seem to respond quickly to landscape change, since no effect of past landscape structure was observed on current species frequencies. Moreover, plant traits did not play a major role in mediating species response to environmental variation in these grasslands. Our findings differ from those obtained in northern and central European grasslands, probably due to differences in methodology but also to the smaller contrast in environmental conditions between grasslands and the adjacent forests and scrub in Mediterranean landscapes.