Interactions between abiotic filters, landscape structure and species traits as determinants of dairy farmland plant diversity

Maintaining farmland biodiversity in Europe under scenarios of agricultural intensification is a keystone challenge of nature conservation. The recruitment of species from the regional pool to local landscape mosaics and individual patches is known to be determined by multi-scale ecological filters. Here we aimed at clarifying the relative importance of the physical environment, land use and landscape structure, and species traits, as filters of landscape-level plant species diversity in intensive farmland. Vascular plant species diversity was surveyed in 18 dairy farmland mosaics along a gradient of agricultural specialisation in Northern Portugal. Plant species were grouped according to their life strategy, biogeographic origin, and synecological preferences. Species richness was found to be highest in lowland areas, where warmer climate and nutrient-rich soils contribute to balance the potential negative effects of intensive farming. Multiple predictors, related to physical environment (e.g. climate), land use (e.g. crop area), and landscape structure (e.g. mean patch size), were found to influence diversity patterns, even under the homogenizing effects of agricultural intensification. Dissimilarity models discriminated distinct types of responses, with patterns for biogeographic and synecological groups of species being better predicted by landscape based models. In contrast, a dominant role of physical predictors was observed in explaining diversity patterns for plant strategies. Overall, our results confirmed that physical environmental gradients, land use, landscape structure, and species traits interact in determining landscape-level plant diversity patterns. Such patterns may influence agro-ecosystem responses to environmental changes, and thus should be considered in the development of agri-environmental policies and monitoring schemes.