Skip to contents

Splits the variation among sites of the community-weighted mean of a trait into the parts due to species turnover, to intraspecific variability and to their covariation, following Lepš et al. (2011), from the trait values of individuals.

Usage

trait_flex(
  individuals,
  traits = "trait",
  site = "site",
  species = "species",
  abundance = NULL,
  formula = ~1,
  env = NULL
)

Arguments

individuals

Data frame with one row per individual.

traits

Names of the columns with trait values.

site, species

Names of the columns with the site and the species.

abundance

Name of a column with the abundance of each population (the same value for all its individuals), or NULL to weight by the number of individuals measured.

formula

One-sided formula with variables of env, with an intercept; ~ 1 (the default) gives the total sums of squares only. With several terms, explained is the sum of squares of the whole model and p the p-value of its F test.

env

Data frame with one row per site, a column named like site and the variables of formula; it is matched to the sites by name. Sites with missing values in these variables are left out of all sums of squares, with a warning.

Value

A list with decomposition, a data frame with one row per trait and component (turnover, intraspecific, covariation, total) and the columns explained, residual, total (sums of squares) and p (of the regression, NA for the covariation and without a model), and sites, a data frame with the specific and fixed means and their difference for each site and trait. The list has class trait_flex, with a plot() method.

Details

At each site, the "specific" mean weights the mean trait of each population (a species at the site) by its abundance, and the "fixed" mean weights the mean of the species over all individuals in the data by the same abundances. Their difference is the part of the community mean due to intraspecific variability. The sums of squares among sites of the specific mean (total), of the fixed mean (turnover) and of the difference (intraspecific) are related by $$SS_{total} = SS_{turnover} + SS_{intraspecific} + SS_{covariation};$$ a positive covariation means that turnover and intraspecific variability move the community mean the same way. With a formula, each of the three site means is regressed on the variables of env, and the sums of squares are split into the parts explained by the model and the residuals, with the p-value of each regression.

Abundances are those of the abundance column when given (one value per population, as in sample_individuals()), otherwise the number of individuals measured. Individuals with a missing trait value are left out of the population means, the species means and the counts of that trait together, so that both community means use the same populations and weights.

References

Lepš, J., de Bello, F., Šmilauer, P., & Doležal, J. (2011). Community trait response to environment: disentangling species turnover vs intraspecific trait variability effects. Ecography, 34, 856–863. doi:10.1111/j.1600-0587.2010.06904.x

Examples

set.seed(1)
ind <- data.frame(site = rep(paste0("s", 1:6), each = 20),
                  species = sample(c("a", "b", "c"), 120, replace = TRUE))
env <- data.frame(site = paste0("s", 1:6), moisture = 1:6)
ind$trait <- c(a = 10, b = 20, c = 30)[ind$species] +
  env$moisture[match(ind$site, env$site)] + rnorm(120)
res <- trait_flex(ind, formula = ~ moisture, env = env)
res$decomposition
#>   trait     component explained   residual     total            p
#> 1 trait      turnover  5.629562 2.43805361  8.067616 3.843592e-02
#> 2 trait intraspecific 17.195890 0.14776307 17.343653 2.729729e-05
#> 3 trait   covariation 19.677940 0.08327175 19.761212           NA
#> 4 trait         total 42.503393 2.66908843 45.172481 1.335888e-03
if (requireNamespace("ggplot2", quietly = TRUE)) plot(res)