Quantifying the impact of landscape and habitat features on biodiversity in cultivated landscapes

Determining habitat and landscape features that lead to patterns of biodiversity in cultivated landscapes is an important step for the assessment of the impact of extensification programmes in agriculture. In the context of an assessment of the effect of national extensification programme on biodiversity in agriculture, field data of three regions (7 km(2) each) were collected according to a stratified sampling method. A distribution model of three taxa (spiders, carabid beetles, and butterflies) is related to influencing factors by means of multivariate statistics (canonical correspondence analysis (CCA), partial CCA). Hypothetical influencing factors are categorised as follows: (1) habitat (habitat type, plant species richness) and (2) landscape (habitat heterogeneity, variability, diversity, proportion of natural and semi-natural areas). The correlative model developed for the spider assemblages revealed that the most important local habitat factors are those directly influenced by management practices. Landscape variability, heterogeneity and diversity in the surroundings are not significant factors. Carabid beetle assemblages show a positive reaction to landscape features and respond to particular cultivated surroundings. The model developed for butterflies shows that species assemblages are sensitive to landscape features. Surrounding land use in particular, has a major influence. There are no general models relating overall species diversity to landscape diversity. The relationship strongly depends on the organism examined. Therefore, biodiversity response to landscape and habitat changes (e.g. the extensification programme) has to be identified by means of multi-taxon concept. (C) 2003 Elsevier Science B.V. All rights reserved.