Skip to contents

Taxonomic Diversity Rarefaction

Source: R/rarefaction_taxonomic.R
rarefaction_taxonomic.Rd

Computes rarefaction curves for taxonomic diversity indices by subsampling individuals from the community at increasing sample sizes. Uses bootstrap resampling to estimate expected diversity and confidence intervals at each sample size.

Usage

rarefaction_taxonomic(
  community,
  tax_tree,
  index = c("shannon", "simpson", "species", "uTO", "TO", "uTO_plus", "TO_plus", "avtd"),
  steps = 20,
  n_boot = 100,
  ci = 0.95,
  seed = NULL,
  correction = c("none", "miller_madow", "grassberger", "chao_shen"),
  parallel = FALSE,
  n_cores = NULL
)

Arguments

community

A named numeric vector of species abundances (integers). Names must match the first column of tax_tree.

tax_tree

A data frame with taxonomic hierarchy. First column is species names, subsequent columns are taxonomic ranks.

index

Which index to rarefy. One of "shannon", "simpson", "uTO", "TO", "uTO_plus", "TO_plus", "avtd", "species" (default: "shannon").

steps

Number of sample-size steps along the curve (default: 20).

n_boot

Number of bootstrap replicates per step (default: 100).

ci

Confidence interval width (default: 0.95).

seed

Optional random seed for reproducibility (default: NULL).

correction

Bias correction for the Shannon index. One of "none" (default), "miller_madow", "grassberger", or "chao_shen". Only used when index = "shannon". Passed to shannon(). See shannon() for details.

parallel

Logical. If TRUE, use parallel processing to speed up bootstrap resampling across sample sizes. Default FALSE.

n_cores

Number of CPU cores to use when parallel = TRUE. Default NULL uses up to 2 cores (CRAN policy limit).

Value

A data frame with columns:

sample_size

Number of individuals in the subsample

mean

Mean index value across bootstrap replicates

lower

Lower bound of the confidence interval

upper

Upper bound of the confidence interval

sd

Standard deviation across replicates

Details

The algorithm works as follows:

  1. Expand the abundance vector into an individual-level vector (e.g., c(sp1=3, sp2=2) becomes c("sp1","sp1","sp1","sp2","sp2")).

  2. For each sample size (from min to total N), draw n_boot random subsamples without replacement.

  3. For each subsample, count species abundances and compute the chosen diversity index.

  4. Return the mean and confidence interval at each step.

References

Gotelli, N.J. & Colwell, R.K. (2001). Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecology Letters, 4, 379-391.

Ozkan, K. (2018). A new proposed measure for estimating taxonomic diversity. Turkish Journal of Forestry, 19(4), 336-346.

See also

plot_rarefaction() for visualising the rarefaction curve, ozkan_pto() for full pTO calculation, shannon() and simpson() for classical indices, avtd() for average taxonomic distinctness.

Examples

comm <- c(sp1 = 10, sp2 = 5, sp3 = 8, sp4 = 2, sp5 = 3)
tax <- data.frame(
  Species = paste0("sp", 1:5),
  Genus   = c("G1", "G1", "G2", "G2", "G3"),
  Family  = c("F1", "F1", "F1", "F2", "F2"),
  stringsAsFactors = FALSE
)
rarefaction_taxonomic(comm, tax, index = "shannon", n_boot = 50)
#> taxdiv -- Rarefaction Curve
#>   Index: shannon 
#>   Total N: 28 
#>   Bootstrap: 50 replicates
#>   CI: 95 %
#>   Steps: 20 
#> 
#>  sample_size      mean     lower     upper        sd
#>            2 0.4990660 0.0000000 0.6931472 0.3143820
#>            3 0.7519484 0.0000000 1.0986123 0.3129182
#>            5 1.0436130 0.5392395 1.5470547 0.2634342
#>            6 1.1578025 0.6365142 1.5607104 0.2258762
#>            7 1.1932895 0.8321740 1.5498260 0.1972654
#>            9 1.2095448 0.8486856 1.5229551 0.1958911
#>           10 1.2659255 0.8979457 1.5453400 0.1943499
#>           12 1.3698280 1.0905674 1.5607104 0.1331104
#>           13 1.3517378 1.1275231 1.5188263 0.1151802
#>           14 1.3991807 1.1635515 1.5457653 0.1019017
#> ... ( 20  rows total)