The index sums each species' share of the surprise
For each species, pi is its share of all individuals (its count divided by the total N). Multiply that share by its natural log, sum across every species, and flip the sign. Rare species pull the value up, dominant ones hold it down. The maximum for S species is H′max = ln(S), reached only when all species are equally common.
Same richness, very different diversity
Two communities can have the exact same number of species yet score far apart, because evenness is what the index really measures.
Both communities hold 100 species. When one species dominates, H′ collapses — richness alone would call them identical.
Reading the number
| H′ (natural log) | Diversity | Typical setting |
|---|---|---|
| 0–0.5 | Very low | Monoculture crop, parking lot |
| 1.0–2.5 | Low–moderate | Grassland, recovering habitat |
| 2.5–3.5 | Moderate–high | Temperate forest |
| 3.5–4.5 | High | Rainforest, coral reef |
| > 4.5 | Very high | Megadiverse regions |
To compare sites of different richness, use Pielou's evenness J = H′ ÷ ln(S), which rescales to 0–1.
Common questions
What is a good Shannon diversity value?
It depends on the habitat. Natural ecosystems usually fall between 2.5 and 4.5, tropical rainforests exceed 4, and monocultures sit near 0. Always compare against a reference site in the same region rather than a fixed number.
Why not just count the species?
A raw species count ignores how common each one is. A stream with 10 species where one makes up 95 percent of individuals is less diverse than a stream with 5 species that are all equally common. The Shannon index captures that balance.
Should I use natural log or log base 10?
Natural log (ln) is the ecology standard and gives values up to about 6. Base 10 works too but produces numbers roughly 2.3 times smaller, so state which base you used when reporting.
