Trait hypervolumes based on natural history collections can detect ecological strategies that are distinct to biogeographic regions

Timothy Harris*, Gianluigi Ottaviani, Mark Mulligan, Neil Brummitt

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)


Abstract: Bioregionalisation partitions diversity so that similarity of the selected biological and ecological variables is higher within regions than it is outside those regions. The classic approach partitions an area based on species composition, whereas more recent methods based on remotely sensed data classify biogeographic regions on biophysical and structural variables of vegetation. Another, yet to be explored opportunity, is offered by identifying distinct ecological strategies of plants inhabiting a given area, that is, a functional trait‐based bioregionalisation. Here, we propose such a bioregionalisation using trait hypervolumes. We also compare the proposed functional bioregionalisation with established classifications based on species composition or on remotely sensed data to identify spatial congruence among them, and suggest possible reasons behind observed patterns. Natural history collections represent an underexploited resource, despite holding both trait and locality information and being taxonomically comprehensive. We compile values of traits (leaf size, plant height, seed number per fruit, seed volume) derived from natural history collections for a random sample of African angiosperm species (~1% of the continental flora) to estimate a trait hypervolume. We use hierarchical clustering to divide the hypervolume into four segments (each representing a distinct ecological strategy), whose spatial intersections produced 12 putative biogeographic regions, each containing one or more of these strategies. We spatially map the hypervolume segments onto the entire African continent and calculate the spatial congruence of the putative functional biogeographic regions with previous bioregionalisations. We identify values and combinations of traits that can be indicative of biogeographic regions. This functional bioregionalisation shows greater spatial congruence with that derived from species composition than from remote sensing. However, spatial congruence is low at the continent scale (19%–37%), and varies greatly among regions and in pairwise comparisons between bioregionalisations. Synthesis. Plant traits from natural history collections offer an underused source of information for biogeographic analyses. We demonstrate potential applications of trait hypervolumes in functional biogeography, and outline strengths and drawbacks of the different bioregionalisation methods. Finally, we suggest that key ecological strategies could be used in future models as proxies to anticipate shifts of species assemblages and biogeographic regions.
Original languageEnglish
Early online date17 Oct 2022
Publication statusE-pub ahead of print - 17 Oct 2022


  • Biogeography
  • Functional ecology
  • Macroecology
  • Africa
  • functional biogeography
  • functional bioregionalisation
  • herbarium collections
  • hypervolume segments
  • leaf‐height‐seed scheme
  • plant trait


Dive into the research topics of 'Trait hypervolumes based on natural history collections can detect ecological strategies that are distinct to biogeographic regions'. Together they form a unique fingerprint.

Cite this