Urban biotic homogenization: approaches and knowledge gaps
Data files
Apr 29, 2022 version files 84.66 KB
-
Lokatis_Jeschke_DataS1.xlsx
83.33 KB
-
README.txt
1.33 KB
Abstract
Urbanization is restructuring ecosystems at an unprecedented pace, with complex and profound consequences for life on earth. One of the hypothesized trajectories of urban ecosystems and species communities is biotic homogenization, possibly leading to very similar species assemblages in cities across the globe. Urbanization can, however, also have the opposite effect: biotic diversification, with cities – at least on the local scale – becoming biologically more diverse, mainly as a consequence of high species introduction rates and habitat diversification. Applying the hierarchy-of-hypotheses approach, we systematically map and structure the comprehensive body of literature on the urban biotic homogenization (UBH) hypothesis, comprising 225 individual studies (i.e. tests of the hypothesis) retrieved from 145 publications. The UBH hypothesis is studied at multiple levels with a multitude of approaches and underlying assumptions. We show that UBH is generally used with two very different connotations: about half of the studies investigated a potential increase in community similarity across cities, whereas the other half investigated biotic homogenization within cities, the latter being supported more frequently. We also found strong research biases: (1) a taxonomic bias towards birds and plants; (2) a bias towards small and medium distances (<5000 km) in comparisons across cities; (3) a dominance of studies substituting space for time vs. true temporal studies; (4) a strong focus on terrestrial vs. aquatic systems; (5) more extraurban (including periurban) areas than natural or rural ecosystems for comparison to urban systems; (6) a bias towards taxonomic vs. functional, phylogenetic and temporal homogenization; and (7) more studies undertaken in Europe and North America than in other continents. The overall level of empirical support for the UBH hypothesis was mixed, with 55% of the studies reporting supporting evidence. Results significantly differed when a natural/nature reserve, an extraurban or rural/agricultural area served as reference to infer biotic homogenization, with homogenization being detected least frequently when urban systems were compared to agricultural, i.e. other anthropogenically influenced, study sites. We provide an evidence map and a bibliographic network, and identify key references on UBH with the goal to enhance accessibility and orientation for future research on this topic.
Methods
We compiled a comprehensive collection of empirical studies addressing the urban biotic homogenization (UBH) hypothesis following the PRISMA protocol (Moher et al., 2009), see Appendix S1: Fig. S3 for PRISMA flowchart. A search within the ISI Web of Science was performed on 9 June 2020 using the following terms: ‘(urban* OR city OR cities) AND (biotic homogenization OR ((biodivers* OR taxonomic diversity OR functional diversity OR genetic diversity OR evolutionary diversity OR species diversity OR trait diversity OR species composition OR habitat diversity) AND homogenization))’.
This search yielded 560 hits which were screened for relevance (abstract and title) and from these, 385 articles were assessed (full-text) and included in our database if (1) the publication addressed the UBH hypothesis or biotic homogenization in cities or in response to urbanization, and (2) biotic homogenization was studied empirically by either direct testing, or inferential reasoning from qualitative evidence. We thus included a wider range of publications than for a formal meta-analysis. We excluded publications if (1) neither quantitative nor qualitative evidence was provided on the UBH hypothesis, or (2) two entries described the same study or presented the same analyses (e.g. doctoral dissertation and subsequent publication of results). For preprints and theses, we checked in Google Scholar and personal profiles of first authors, e.g. on ResearchGate, whether a more recent publication of the results in an English journal was available and included it instead if available. We also searched manually whether non-English and non-German publications that were retained after the first screening had since been published in English, or else excluded them from the analysis. 145 articles fulfilled these criteria and were thus included in our dataset.