3D vegetation structure influence on boreal forest bird species richness

Abstract: Bird populations across the world are in danger with decreasing numbers and more species continually becoming red-listed. One main driver behind this trend is human-caused habitat loss and degeneration, which in particular has been identified as a major threat in forested regions. The importance of forest vegetation structure for bird diversity has been shown in many studies, though typically for small restricted study areas. Here I used a large region of interior boreal Sweden as study area. I used point census count data from the Swedish National Bird Monitoring program combined with recently published nation-covering lidar data, to investigate how bird species richness was affected by 3D forest structure. In total 37 forest-associated bird species were included. Non-parametric random forest models and generalized linear models (GLMs) were used, rendering R2 values of 36% and 15%, respectively. Variation in vegetation density and canopy height were the two most important forest structure features to predict bird species richness. Height evenness, also known as foliage height diversity (FHD), scored low in variable importance despite being considered a significant driver of bird diversity by many authors. A constrained correspondence analysis (CCA) ordination method was performed to explore habitat selection and niche width for individual bird species. Species with similar habitat preferences were nested in the CCA diagram but showed large overlaps, suggesting that there is a signal in the data but also much noise. Thus, separating between habitat generalists and specialists was not possible. For conservation applications and prioritizations, bird species richness is not necessarily a suitable measure. Rather, the contribution to beta and gamma diversity, as well as the specific habitat preferences of rare, red-listed and specialist species, should guide conservation measures and forest management practices. Future studies should extend further towards a landscape-based study design where forest fragmentation and configuration are significant components.