Understanding the conditions under which species traits, species-environment relationships, and the spatial structure of the landscape interact to shape local communities requires quantifying the relative contributions of space and the environment on community composition. Using analogous sampling of arboreal and terrestrial oribatid mite communities across a large spatial scale in a temperate rainforest, we quantified the variation in oribatid mite community structure relating to environmental and spatial factors, and tested whether terrestrial and arboreal communities demonstrated a difference in their patterns of community composition based on the assumption of differences in dispersal potential. The expectation that terrestrial oribatid mite communities are spatially structured while arboreal communities are environmentally structured was supported by our analyses at the level of variation in beta diversity, but not by assessing beta diversity itself. We found that terrestrial oribatid mite communities with active, cursorial dispersal demonstrate spatial constraint consistent with reduced long-distance dispersal opportunities and high environmental dissimilarity among sites. Arboreal communities, which potentially disperse long distances via passive aerial vectors, show a spatial signature associated with patterns in beta diversity and a correlation with environmental dissimilarities among sites. In the arboreal community, moisture content of the substrate, total tree height, and average sampled branch height were significant factors explaining beta diversity patterns. For ground-dwelling species, predator abundance and soil type were important local determinants of community variability. Both communities showed clear spatial structuring, suggesting that dispersal limitation continues to influence community composition across multiple forest watershed locations. Our results provide evidence of dispersal-maintained diversity patterns in response to local environmental factors in arboreal and terrestrial communities. The relative importance of stochastic dispersal assembly may be dependent on strong deterministic effects associated with micro-site and macro-site environmental variation, particularly across large spatial scales.