Association of tree and plot characteristics with microhabitat formation in European beech and Douglas-fir forests.

Process-orientated, unmanaged forest remnants are not sufficient for halting the loss of forest biodiversity. Thus, integrated biodiversity-promoting management for forest inhabitants is needed. Microhabitats, such as tree cavities or bark pockets, are essential for the preservation of saproxylic species and of critical importance for endangered ones. This study investigates (1) which factors trigger the formation of microhabitats at both the individual tree and aggregated plot level, and (2) whether the co-occurrence of microhabitats differs between managed (=logged) and unmanaged forests. Relationships between the occurrence of 17 microhabitat types and individual tree features (e.g. light availability, and tree vitality) and plot characteristics (e.g. stand density index and stand age) in 398 plots dominated by Fagus sylvatica or Pseudotsuga menziesii in Germany and the USA were studied using random-effects logistic and normal regression modelling. Separate analyses were performed for German beech forests, German Douglas-fir forests, and the US Douglas-fir forests. Our results show that (1) tree diameter in breast height (DBH), tree vitality and branchiness or epicormic branches are highly related with the occurrence of one or more microhabitats on individual trees in managed and unmanaged beech and US Douglas-fir forests. In managed German Douglas-fir forests, vitality is not a predictor for the occurrence of microhabitats on a tree, but tree density and the maximum age of trees in a stand in addition to DBH and branchiness have an effect. Time since last management is not a statistically significant predictor for the presence of microhabitats at the tree level, but it is for German beech at the plot level. In Douglas-fir-dominated forests both in Germany and in the USA, the stand density index was the only common predictor at the plot level. (2) Unmanaged German beech and Douglas-fir forests exhibit more statistically significant and positive correlations with microhabitat groups than managed stands, implying that the presence of one microhabitat group on a tree is associated with the presence of other microhabitat groups. We finally conclude that measures for supporting microhabitat inhabitants in managed forests are scale and species dependent (tree versus plot level; beech versus Douglas-fir-dominated forests). Trees that carry microhabitats seem to have similar features independently of forest management. At the plot level, density management may trigger the accumulation of microhabitats. Our results indicate that in forest management, it is possible to consider the factors influencing the formation of microhabitats and implement adequate forest practices to advance their formation. [ABSTRACT FROM AUTHOR]