Forest management practices and forest fragmentation are commonly recognised as factors influencing species distribution and diversity, but it remains largely unknown how species turnover and fluctuations across years might affect generalisations from short-term predictions. There is also much uncertainty about how compositional heterogeneity within forests, through their partitioning into differently managed stands, influences species occurrence and diversity. We aimed to investigate bird species richness in different forest stands in a heterogeneous landscape and asked whether short-term studies commonly used are sufficient to describe diversity–habitat relationships.

Breeding birds were monitored in 50 stands of differing forest types and management regimes in 2008 and 2009 in the heterogeneous forest production landscape of the Schwäbische Alb in SW Germany. Based on a database of 25 and 21 species observed in 2008 and 2009, respectively, we estimated local species richness and species turnover with Bayesian hierarchical multi-species occupancy models to account for imperfect detection. A suite of forest attributes at stand level (from local field inventories) and landscape level, between 100 and 2000 m in extent, which were derived from forest management maps, were screened for their explanatory power regarding bird species richness. The most meaningful variables were included in a final set of models to account for effects of forest attributes on species occurrence and resulting gradients in species richness and turnover.

Bird species richness in 2008 was mostly explained by stand composition heterogeneity over a distance of 800 m around sample locations, whereas in 2009 local stand features like stand succession type and stand age mostly explained species richness. Species turnover was significantly higher in young forest stands of thicket and pole-wood stages compared with old forest stands of timber stages.

Variable multi-scale impacts of forest attributes on local species richness and turnover, together with year-to-year fluctuations in the regional species pool, emphasise that the relative importance of local versus landscape-scale factors in determining bird species richness may vary across years. Long-term and multiple-scale investigations that take structural and compositional forest heterogeneity into account are necessary for improved prediction of species–habitat relationships.