Search for a command to run...
Abstract Aim To investigate the performance and relative importance of abiotic and biotic predictors of species richness of three taxa in forest‐dominated landscapes across an environmentally heterogeneous mountain region. Location Switzerland (central Europe). Methods We used a broad set of nationally available environmental predictors grouped into (1) climate, (2) topography and soil and (3) 3‐D vegetation structure derived from airborne Light Detection and Ranging (Li DAR ) data to spatially predict the forest species richness of vascular plants, butterflies and breeding birds. We used presence data of 212 plant, 157 butterfly and 92 bird species from multiple transect samples in > 220 1 km 2 squares at elevations between 261 and 2123 m a.s.l. across 41,248 km 2 . We applied an ensemble modelling approach consisting of five modelling techniques and evaluated their predictive performance using the cross‐validated percentage of explained variance of each predictor group separately and the combinations thereof. We investigated the relative importance and response of each predictor and partitioned the variation into independent and shared components per variable group. Results Climate performed best in predicting forest species richness across taxa. Vegetation structure particularly improved the predictions of butterfly and bird species richness, while soil pH was an important predictor for forest plant species richness. Climate appeared to be mainly indirectly related to butterfly species richness, via correlations with habitat type and structure. The strength and direction of the relationships between the predictors and species richness were taxon‐specific with low cross‐taxon congruence. Main conclusions The growing availability of Li DAR data offers powerful new tools for describing vegetation structure and associated animal habitat quality across large areas. This will further our understanding of niche‐driven assembly processes in forest landscapes. Although climate was the dominant factor controlling species richness across taxa from different trophic levels, the taxon‐specific distributional pattern and response to environmental conditions emphasize the difficulty of accounting for a range of taxa in prioritising biodiversity conservation measures.