Comparative influence of spatial scale on beta diversity within regional assemblages of birds and butterflies

Aim We examined whether variation in species composition of breeding birds and resident butterflies in the Great Basin of North America depended on sampling grain (the smallest resolvable unit of study) and on the relative proximity of sampling units across the landscape. We also compared patterns between the two taxonomic groups with reference to their life‐history characteristics. Location Data for our analyses were collected from 1996 to 2003 in three adjacent mountain ranges in the central Great Basin (Lander and Nye counties, Nevada, USA): the Shoshone Mountains, Toiyabe Range and Toquima Range. Methods Data on species composition for both taxonomic groups were collecting using standard inventory methods for birds and butterflies in temperate regions. Data were compiled at three sampling grains, sites (average 12 ha), canyons (average 74 ha) and mountain ranges. For each sampling grain in turn, we calculated similarity of species composition using the Jaccard index. First, we investigated whether mean similarity of species composition among the three ranges differed as a function of the grain size at which data were compiled. Secondly, we explored whether mean similarity of species composition was greater for canyons within the same mountain range than for canyons within different mountain ranges. Thirdly, we examined whether mean similarity of species composition at the site level was different for sites within the same canyon, sites within different canyons in the same mountain range, and sites within canyons in different mountain ranges. We used a Bayesian model to analyse these comparisons. Results For both taxonomic groups, mean similarity of species composition increased as the sampling grain increased. The effect of spatial grain was somewhat greater for birds than for butterflies, especially when the intermediate sampling grain was compared with the smallest sampling grain. Similarity of species composition of butterflies at each sampling grain was greater than similarity of species composition of birds at the same grain. Mean similarity of species composition of both birds and butterflies at the canyon level and site level was affected by relative proximity of sampling locations; beta diversity increased as the relative isolation of sampling locations increased. Main conclusions The sensitivity of beta diversity to sampling grain likely reflects the effect of local environmental heterogeneity: as sampling grain increases, biotic assemblages appear more homogeneous. Although breeding birds in our study system have larger home ranges than resident butterflies, birds may have more specialized resource requirements related to vegetation structure and composition, especially at small sampling scales. The degree of variation in species composition of both taxonomic groups suggests that spatially extensive sampling will be more effective for drawing inferences about regional patterns of species diversity than intensive sampling at relatively few, smaller sites.