Table 1.
Overview of species richness estimators. Column “up” indicates whether an estimator is an upsampling estimator (
) or not (
; then it is a population estimator). Population (asymptotic) and upsampling (extrapolating) estimators are separated by a horizontal line. Column “par” indicates whether an estimator is parametric (
), i.e. whether it makes distributional assumptions about the species composition, or not (
). 
The Chao 1 estimator can be derived from a Poisson model, but was first introduced as a nonparametric estimator. Column “Implementation” links to the estimator’s implementation that we used for computational experiments (“ours” means our own implementation; available at https://gitlab.com/rahmannlab/speciesrichness).
| Name | Up | Par | Reference | Implementation |
|---|---|---|---|---|
| Good–Turing |
|
|
Good [2] | ours |
| Jackknife |
|
|
Burnham and Overton [8] | ours |
| ACE |
|
|
Chao and Lee Chao and Lee [16] | ours |
| Poisson |
|
|
Sandland and Cormack [9] | Breakaway R package |
| Chao 1 |
|
|
Chao [7] | ours |
| Gamma-Poisson mixture |
|
|
Fisher et al. [1] | ours |
| Chao and Bunge |
|
|
Chao and Bunge [17] | Breakaway R package |
| Lanumteang and Böhning |
|
|
Lanumteang and Böhning [18] | ours |
| Chiu |
|
|
Chiu [19] | ours |
| Objective Bayesian |
|
|
Barger and Bunge [20] | Breakaway R package |
| Recon |
|
|
Kaplinsky and Arnaout [14] | GitHub ArnaoutLab |
| Valiant |
|
|
Valiant and Valiant [11] | Valiant Code |
| Breakaway |
|
|
Willis and Bunge [12] | Breakaway R package |
| TES |
|
|
Zou et al. [21] | TES R script |
| iNEXT |
|
|
Hsieh et al. [22] | iNEXT R package |
| Smoothed Good–Toulmin |
|
|
Orlitsky et al. [23] | ours |
| PreSeq |
|
|
Daley and Smith [24] | PreSeq R package |
| Pitman sampling formula |
|
|
Pitman [25] | GitHub Stefanie Tauber |
| DivE |
|
|
Laydon et al. [26] | DivE R package |
| RichnEst (formerly Dupre) |
|
|
Schröder and Rahmann [10] | GitLab RahmannLab |