Scale-independence in beetle species diversity-productivity relationship: The importance of heterogeneity arrangement in space

Species diversity-productivity relationship is often, but not always, scale dependent. We hypothesize that the arrangement of environmental heterogeneity along a rainfall gradient, as a surrogate for productivity, determines whether scale dependence of beetle diversity at our sandy sites exists. We chose seven geographically distant landscapes (360 kms from south to north) within a mean annual rainfall range of 30–585 mm, each having six clusters of five uniformly distributed pitfall traps (within a range of 0.24 km2). Vegetation and physical variables were measured within each landscape and its clusters to characterize both spatial and temporal heterogeneity. We used species diversity of each cluster and the accumulative species diversity in a landscape as local and regional diversity, respectively. Between-clusters dissimilarity in diversity reflected the turnover of species within a landscape. We used environmental measures in each cluster and at the entire landscape to indicate local and regional heterogeneity, respectively, and changes between transects to characterize the within-landscape dissimilarity in heterogeneity. Beetle diversity and productivity showed a unimodal pattern at both spatial scales, with a peak at intermediate productivity of ca. 300 mm. Environmental heterogeneity increased with productivity both at the local and at the landscape scales, but within-landscape dissimilarity in heterogeneity did not change with productivity. Between-clusters dissimilarity in diversity was higher at intermediate productivity, explaining the unimodal pattern at the landscape scale. We suggest that the scale independence in our study results from the equal distribution of environmental heterogeneity within a landscape. We propose that the exploration of the diversity-productivity relationship across scales should consider the way heterogeneity is distributed along the productivity gradient.