Spatial structure and scaling of beetle, bird and plant communities in North American grasslands

The spatial structure and scale of biotic communities are functions of the natural history of the organisms and the heterogeneity of the landscape in which these organisms exist. The relationships between these organisms, the environment and geographic space differ at different locations. The primary goal of this dissertation was to present new methods by which to understand these relationships. To examine these relationships, ground-dwelling beetles, breeding birds, and vascular plant communities were studied at four sites along the transition from short-grass steppe in Colorado to tallgrass prairie in Kansas. At these four sites, two 2-km transects were established along which taxonomic surveys were conducted and environmental data, including soil texture, shrub density and landsat imagery, were collected in the late spring of 2000 and 2001. Environmental data were collected at three scales: trap scale (1 m2), local scale (75--300 m2) and landscape scale (900 m2). Using these geographically referenced environmental and community data, three questions were addressed: (1) How are different biotic communities structured with respect to their mobility and response to environmental heterogeneity? (2) What environmental and spatial factors explain the variance in this spatial structure? and (3) Are the patterns of species richness within these communities predictable based on area and heterogeneity? To answer these questions I employed three different methods. Multi-scale ordination with canonical correspondence analysis was used to determine the spatial structure of these communities. Using hierarchical variance partitioning, the variance in community structure was apportioned by scale to assess the ability of environmental variables measured at different scales to explain community structure. Species-area-habitat relationships were quantified using a scale-dependent measure of heterogeneity, which enabled the predictions of species accumulation curves across the study region. Based on these analyses, it was shown that there are systematic differences in community structure across a regional gradient, there are differences between taxonomic groups at specific sites and across the region that are not consistent and underscore the limitations of using "indicator" taxa to predict overall biodiversity, and the three methods used in this dissertation highlight current progress in understanding scale-dependent species-environment relationships.