Predicting the species richness of Alpine pastures using indicator species

1 Plant species richness is often used as a criterion in the assessment of ecological quality. Complete species inventories are time-consuming to establish, and the resulting infor-mation may be more detailed than needed. In this study we investigated how reliably species richness on Alpine pastures can be estimated from the presence or absence of a limited number of indicator species. 2 Based on data from a vegetation survey in 200 plots (1 m2) on ten Alpine pastures in Glarus (northern Swiss Alps), we identified 36 vascular plant species that were signifi-cantly more frequent in species-rich than in species-poor plots. The number of these ‘richness indicators ’ in a relevé increased linearly with actual plant species richness (r2 = 0.80) but was only weakly related to the number of non-indicator plant species (r2 = 0.10). 3 If subsets of 3, 5, 8, 10 or 20 species are randomly drawn from the 36 ‘richness indicators’, the precision of species richness estimates increases with increasing subset size and with increasing frequency of the species in the subset. For the most species-rich relevés, the precision of species richness estimates is already close to maximum with subsets of eight species. 4 Character species from species-rich Alpine pastures at regional scale (phytosocio-logical classification) are found to be less effective in predicting species richness than our locally defined set of ‘richness indicators’. 5 Pasture areas with high species richness ( ≥ 25 plant species m-2) can be roughly identified from the presence of at least half of the species in a given indicator species set (which should consist of at least eight species). This information can be used for a rapid mapping of the most species-rich parts within Alpine pastures.