Species Accumulation Curve — Ecological Sampling Simulator

Landscape

Species pool 16

Spatial clustering 0.60

Individuals per plot 10

Adjust the sliders, then press "New landscape" to regenerate — mid-survey changes would silently rewrite the community you're sampling.

Survey

Live Readouts

Plots sampled

0 / 64

True species pool

Species observed (S_obs)

Chao2 estimate (Ŝ)

0.0

Observed S(n) Chao2 Ŝ True pool

How it works

A species-accumulation curve S(n) plots cumulative species richness observed against the number of sampling units (here, quadrat plots) surveyed. It always rises and decelerates — early plots turn up mostly new species, later ones mostly re-find species you already have.

S(n) = number of distinct species seen in the first n plots
f1 = "uniques"   = species found in exactly 1 sampled plot
f2 = "duplicates" = species found in exactly 2 sampled plots

Chao2 (incidence-based) estimator:
  Ŝ = S(n) + f1(f1 - 1) / (2(f2 + 1))

Chao2 corrects the raw count S(n) for species that exist in the landscape but haven't been detected yet. Its logic: if many species have turned up in only one plot so far (f1 large), that's a strong signal more undetected species remain; if almost nothing is a "unique" any more, the survey is close to complete and Ŝ ≈ S(n).

Species pool — how many species truly exist across the whole 8×8 landscape (the number Ŝ should converge toward).
Spatial clustering — 0 spreads every species near-uniformly across all 64 plots (a few random plots find almost everything); 1 concentrates each species tightly around one "home" plot, so plots must be spread out to catch rare, patchily-distributed species — the realistic case for most field surveys.
Individuals per plot — local sample size; too few individuals per plot under-detects rare local species even after many plots.
Sampling order — Random draws plots in a random permutation; Systematic sweeps row by row; Clustered grows outward from one seed plot (a "convenience sample"), typically inflating apparent uniques and biasing Ŝ because nearby plots are spatially correlated, not independent.

Real-world relevance: this is the same statistic ecologists use to decide "how many more transects/traps/plots do we need before we've found essentially every species here?" — used in biodiversity assessments, environmental-impact surveys and camera-trap studies (Chao 1984, 1987; Colwell & Coddington 1994).

Green pads = sampled plots. Cones = individuals found, colored by species. Grey pads are still unsurveyed.

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.