Skip to contents

The concept

Species counts treat every species as equally different. But a forest of ten closely related oaks differs from one with ten trees from ten families, and a community of ten species with identical traits performs one ecological role. Two further dimensions capture this:

Dimension Weighs species by Question it answers
Taxonomic diversity (TD) Nothing: all species equally distinct How many species?
Phylogenetic diversity (PD) Branch lengths of a phylogeny (Chao, Chiu and Jost 2010) How much evolutionary history?
Functional diversity (FD) Distances between species’ traits (Chao et al. 2019) How many distinct ecological roles?

tidydiversity measures all three as Hill numbers, so they are in the same unit, the effective number of species, lineages or functional groups, and can be compared directly:

  • PD (with pd_unit = "mean") is the effective number of equally distinct lineages spanning the tree’s depth.
  • FD treats species closer than a threshold τ\tau in trait space as functionally identical, and counts the effective number of distinct functional groups. By default τ\tau is the mean distance between two individuals; tau = "auc" integrates over all thresholds so none has to be chosen.

When FD or PD is much lower than TD, many species are redundant: they duplicate traits or lineages already present.

Data

Trees at the edge and in the interior of a Brazilian Atlantic rainforest fragment, with a dated phylogeny and a trait distance matrix, shipped with the iNEXT.3D package (Chao et al. 2021).

library(tidydiversity)
library(dplyr)
library(ggplot2)

data(list = c(
  "Brazil_rainforest_abun_data", "Brazil_rainforest_phylo_tree",
  "Brazil_rainforest_distance_matrix"
), package = "iNEXT.3D")
wide <- Brazil_rainforest_abun_data
brazil <- tibble(
  site = rep(colnames(wide), each = nrow(wide)),
  species = rep(rownames(wide), ncol(wide)),
  n = unlist(wide, use.names = FALSE)
) |> filter(n > 0)
tree <- Brazil_rainforest_phylo_tree
traits <- as.matrix(Brazil_rainforest_distance_matrix)
count(brazil, site, wt = n, name = "trees")
#> # A tibble: 2 × 2
#>   site     trees
#>   <chr>    <int>
#> 1 Edge      1794
#> 2 Interior  2074

Three dimensions at equal coverage

As in Chapter 3, compare sites at equal sample coverage.

std <- div_standardise(brazil,
  abundance = n, type = "all", tree = tree,
  distance = traits, coverage = 0.97, nboot = 0
)
tab <- std |>
  select(type, site, q, diversity) |>
  tidyr::pivot_wider(names_from = site, values_from = diversity)
knitr::kable(tab, digits = 1,
  caption = "Taxonomic (TD), phylogenetic (PD) and functional (FD) diversity at 97 % coverage."
)
Taxonomic (TD), phylogenetic (PD) and functional (FD) diversity at 97 % coverage.
type q Edge Interior
TD 0 383.3 433.8
TD 1 166.8 175.7
TD 2 83.7 73.3
PD 0 70.9 79.5
PD 1 5.3 5.8
PD 2 1.8 1.9
FD 0 6.9 5.9
FD 1 6.5 5.2
FD 2 6.3 4.7
ggplot(std, aes(q, diversity, colour = site)) +
  geom_line(linewidth = 0.9) +
  geom_point() +
  facet_wrap(~type, scales = "free_y") +
  scale_colour_tidydiversity() +
  labs(y = "Effective number", colour = NULL) +
  theme_tidydiversity()

Diversity profiles at 97 percent coverage for TD, PD and FD at the forest edge and interior

How to read it.

  • Compare the two sites within each panel: which is more diverse in species, lineages and functions.
  • Compare the size of numbers across panels: TD is in species, PD in lineages and FD in functional groups. FD values far below TD mean many species are functionally redundant.
  • Compare the shape of each profile: a steep fall from q=0q = 0 to q=2q = 2 means a few species, lineages or functions dominate.

Interpretation. The three dimensions tell different stories:

  • Species (TD): the interior has more species than the edge at q=0q = 0 (433.8 against 383.3), but the edge has more dominant species at q=2q = 2 (83.7 against 73.3): the interior is richer but more dominated by a few abundant species.
  • Lineages (PD): the interior holds slightly more evolutionary history (79.5 against 70.9 effective lineages at q=0q = 0).
  • Functions (FD): the edge is functionally more diverse at every order (6.9 against 5.9 functional groups at q=0q = 0), even though it has fewer species. A plausible explanation, to be tested rather than assumed, is that edge conditions (more light, drier air, disturbance) admit species with a wider range of traits, while the interior’s extra species are largely functionally redundant.

All FD and PD values are far below the species numbers: many species share lineages and traits. So the interior’s higher species count does not mean more ecological roles, which is why reporting TD alone can mislead conservation priorities.

Choosing the threshold for FD

tau Effect
Small Only near-identical species merge; FD approaches TD
Default (mean distance) Species closer than average are treated as functionally equivalent
Large Most species merge; FD approaches 1
"auc" Integrates over every threshold from 0 to 1: no choice needed; recommended for comparisons

Further methods

The traits article covers community-weighted means, classic FD indices (FRic, FEve, FDiv, FDis, Rao’s Q), null models, fourth-corner analysis, intraspecific variation and trait probability densities. The dimensions article shows incidence data and rarefaction curves for all three dimensions.

How to report it

Taxonomic, phylogenetic and functional diversity were quantified as Hill numbers (q = 0, 1, 2) at 97 % sample coverage (Chao et al. 2021) with tidydiversity, whose estimators reproduce iNEXT.3D. Phylogenetic diversity used the dated phylogeny (effective number of lineages); functional diversity used Gower trait distances with the default threshold (mean pairwise distance; Chao et al. 2019).

References

Chao, A., Chiu, C.-H. and Jost, L. (2010). Phylogenetic diversity measures based on Hill numbers. Philosophical Transactions of the Royal Society B, 365, 3599-3609.

Chao, A. et al. (2019). An attribute-diversity approach to functional diversity, functional beta diversity, and related (dis)similarity measures. Ecological Monographs, 89, e01343.

Chao, A., Henderson, P. A., Chiu, C.-H., Moyes, F., Hu, K.-H., Dornelas, M. and Magurran, A. E. (2021). Measuring temporal change in alpha diversity: a framework integrating taxonomic, phylogenetic and functional diversity and the iNEXT.3D standardization. Methods in Ecology and Evolution, 12, 1926-1940.