A meta-analysis of the effects of urbanization on ground beetle communities

Urbanization is one of the most disruptive forms of land use, and its effects on ecological communities are likely to increase as the human population continues to shift from rural to urban living. Whether the myriad changes encompassed by urbanization (e.g., habitat fragmentation, the urban heat island effect, disturbance) affect biological communities in predictable ways remains unresolved. In this study, we employed meta-analysis to quantify the overall effects of urbanization on ground beetles (Coleoptera: Carabidae), an ecologically diverse and important group that has been relatively well-studied in cities. We calculated effect sizes for carabid species richness, the total abundance of beetles at a site (assemblage abundance), and the abundances of individual species at urban compared to rural forested sites. We found that the number of carabid species in cities was only 77.6% that of corresponding rural sites, whereas assemblage abundance was relatively consistent between rural and urban sites. Such assemblage-level patterns resulted from the loss of certain species from urban sites and the increase in abundance of others. We hypothesized that differences among species would be largely attributable to their ecological traits (body size, habitat affinity, dispersal ability, trophic position). Through model selection based on information theoretic measures, we found support for the importance of the simple and interactive effects of beetle size, habitat affinity, and trophic position, with the most negative effect sizes for large, forest-specialist beetles that were predatory or omnivorous. Although species traits were key to understanding variation in species abundance effect sizes, several remaining sources of variation in carabid beetle responses to urbanization remain. Specifically, although many species exhibited consistent responses across the studies in which they were sampled and might therefore be considered urban avoiders or urban exploiters, many others were either relatively insensitive to the urbanization gradients or exhibited widely disparate responses across studies. Furthermore, we were unable to attribute variation in species richness and assemblage abundance effect sizes to available measures such as city size and study duration. Instead, additional mechanistic studies across the urbanization gradient will be necessary to understand such variation among studies and for species exhibiting variable responses.