Plot-based butterfly surveys: statistical and methodological aspects

Butterflies (Lepidoptera Rhopalocera) are considered good indicators of Europe's biodiversity, especially for grasslands and other open habitats. Therefore monitoring butterfly communities is a powerful tool to measure biodiversity patterns in relation to pressures and to define management actions aimed at conserving and/or increasing biodiversity. However, the precision and efficiency of the methods used to assess the diversity of butterflies have seldom been examined. A mathematical exploration of the correlation between sampling effort and the resulting exhaustiveness of butterfly surveys should allow to select methodological parameters rationally and to estimate the precision of the results obtained. Optimization of these parameters for a given amount of total time should be done with statistical techniques. The aim of this paper is to provide data to estimate the influence of some of the above mentioned parameters on the exhaustiveness of butterfly data for plot-based surveys. Species-rarefaction curves are used to investigate the influence of the duration of single surveys and of repeated visits to a site on the species richness ascertained. Additionally, correlations of the number of days between two butterfly surveys and community dissimilarities are presented. It was found that species richness increased with the number of surveys but this increase differed between lowland and mountain sites and a correlation of the number of days between two butterfly surveys and community similarity was found for the mountain sites, but not for the lowland plots. The differences between mountain and lowland communities are likely related to differences in voltinism in the species present locally.