Canopy structure influences arthropod communities within and beyond tree identity effects: Insights from combining LiDAR data, insecticidal fogging and machine learning regression modelling

•Canopy habitats require more attention in times of global biodiversity loss.•Using LiDAR and canopy fogging, we link canopy structure and arthropod communities.•We use machine learning regression to select variables driving arthropod communities.•Canopy structure drives arthropod communities stronger than tree species identity.•Structural complexity and variability of inter and intra-canopy gaps are key drivers. Forest canopies host an abundant but often neglected diversity of arthropods, which requires careful attention in times of ongoing biodiversity loss. Yet, how tree species composition interacts with canopy structure in shaping arthropod communities remains largely unknown. Here, by combining mobile laser scanning and insecticidal fogging with a machine learning algorithm, we studied which canopy architectural properties affect canopy arthropod communities in monospecific and mixed stands of broadleaved European beech and the coniferous Norway spruce and non-native Douglas fir in Germany. Evaluating the abundances and ecological guild diversity of ∼ 90,000 arthropods and 27 partly novel high-resolution structural variables, we identified vegetation volume and tree species identity as weak predictors of arthropod abundance and ecological guild diversity. In contrast, structural heterogeneity, i.e. structural complexity, vertical layering and variability of canopy gaps—which were highest in coniferous stands—were strong positive drivers. Despite this, arthropod ecological guild diversity was lower in non-native Douglas fir. Mixed stands had intermediate arthropod abundance and ecological guild diversity. Our study shows that habitat heterogeneity and tree species-identity are closely interlinked in shaping associated canopy arthropod communities. Positive effects of habitat heterogeneity on arthropod ecological guild diversity were often uncoupled from resource availability, and the key role of our novel intra-canopy gap indices suggests that they should be considered as indicators in future research on forest heterogeneity-diversity relationships. Broadleaf-conifer mixtures may be suitable to mediate negative tree-species identity effects when adapting forests to global change.