Different facets of tree sapling diversity influence browsing intensity by deer dependent on spatial scale

Browsing of tree saplings by deer hampers forest regeneration in mixed forests across Europe and North America. It is well known that tree species are differentially affected by deer browsing, but little is known about how different facets of diversity, such as species richness, identity, and composition, affect browsing intensity at different spatial scales. Using forest inventory data from the Hainich National Park, a mixed deciduous forest in central Germany, we applied a hierarchical approach to model the browsing probability of patches (regional scale) as well as the species-specific proportion of saplings browsed within patches (patch scale). We found that, at the regional scale, the probability that a patch was browsed increased with certain species composition, namely with low abundance of European beech (Fagus sylvatica L.) and high abundance of European ash (Fraxinus excelsior L.), whereas at the patch scale, the proportion of saplings browsed per species was mainly determined by the species’ identity, providing a “preference ranking” of the 11 tree species under study. Interestingly, at the regional scale, species-rich patches were more likely to be browsed; however, at the patch scale, species-rich patches showed a lower proportion of saplings per species browsed. Presumably, diverse patches attract deer, but satisfy nutritional needs faster, such that fewer saplings need to be browsed. Some forest stand parameters, such as more open canopies, increased the browsing intensity at either scale. By showing the effects that various facets of diversity, as well as environmental parameters, exerted on browsing intensity at the regional as well as patch scale, our study advances the understanding of mammalian herbivore–plant interactions across scales. Our results also indicate which regeneration patches and species are (least) prone to browsing and show the importance of different facets of diversity for the prediction and management of browsing intensity and regeneration dynamics