Non-native ants are breaking down biogeographic boundaries and homogenizing community assemblages

Abstract

As geographic distance increases, species assemblages become more distinct, defining global biogeographic realms with abrupt biogeographic boundaries. Yet, it remains largely unknown to what extent these realms may change because of human-mediated dispersal of species. Focusing on the distributions of 309 non-native ant species, we show that historical biogeographic patterns have already broken down into tropical versus non-tropical regions. Importantly, we demonstrate that these profound changes are not limited to the distribution patterns of non-native ants but fundamentally alter biogeographic boundaries of all ant biodiversity (13,774 species). In total, 52% of ant assemblages have become more similar, supporting a global trend of biotic homogenization. Strikingly, this trend was strongest on islands and in the tropics, which harbor some of the most vulnerable ecosystems. Overall, we show that the pervasive anthropogenic impacts on biodiversity override biogeographic patterns resulting from millions of years of evolution, and disproportionally affect particular regions.

Fig. 1

Conceptual diagram of analysis steps, data flow, and biogeographical analysis.

Fig. 2. Global biogeographic patterns before and after human-mediated dispersal of non-native species.

Biogeographic groups of 309 non-native ant species: a before (native ranges) and b after human-mediated dispersal (native + non-native ranges). Biogeographic realms of all ant species (13,774): c before (native ranges) and d after human-mediated dispersal of non-native species (native + non-native ranges). Colors indicate biogeographic clusters identified using compositional dissimilarity (${\beta }_{\mathrm{sim}}$ index) and clustering analysis. Light gray areas highlight polygons for which no distributional data were available. Source data are provided as a Source Data file.

Fig. 3. Changes in biogeographic boundaries split by mainlands and islands.

Biogeographic realms of all ant species (13,774) on mainlands (top) and islands (bottom): a–c before (native ranges) and b–d after human-mediated dispersal (native + non-native ranges) of non-native species. Colors indicate realms identified using compositional dissimilarity (${\beta }_{\mathrm{sim}}$ index) and clustering analysis. Source data are provided as a Source Data file.

Fig. 4. Global homogenization trend for all ant species (13,774 species) at the global scale.

a Homogenization by polygon (${\overline{h}}_{i\bullet }$) grouped by categories of climatic status (tropical vs. non-tropical) and geographic status (mainlands vs. islands). b Average degree of homogenization (${\overline{h}}_{i\bullet }$) by polygon. c Number of donated and received non-native species grouped by categories of climatic status (tropical vs. non-tropical) and geographic status (mainlands vs. islands). Differences between groups were tested using a two-sided Wilcoxon signed-rank test with Bonferroni correction, and significance are given with asterisk (ns non-significant, *: 0.05 < p ≤ 0.1, **: 0.01 < p ≤ 0.05, ***: p ≤ 0.001). Box plots (a–c) represent data from n = 536 polygons (68 non-tropical islands, 232 non-tropical mainlands, 84 tropical islands, and 152 tropical mainlands) where the lower bound of lower whisker shows the minimum value of the data that is within 1.5 times the interquartile range under the 25th percentile, lower bound of box shows the lower quartile, center of box shows the median, upper bound of box shows the upper quartile, and upper bound of upper whisker shows the maximum value of the data that is within 1.5 times the interquartile range over the 75th percentile. Source data are provided as a Source Data file.

Fig. 5. Global distance-decay before and after human-mediated dispersal across all ant assemblages (13,774 species).

Distance-decay represents the relationship between compositional similarities (1−${\beta }_{\mathrm{sim}}$) of ant assemblages (for all ant species) and geographical distances between the centroids of polygons before (a) and after (b) the human-mediated dispersal of 309 non-native ant species. Compositional similarities are fitted against distance with nls models. The R-square of the models and distance-decay rate per 1000 km are given above the graphs. Source data are provided as a Source data file.