Environmental Engineering Reference
In-Depth Information
Simpson's index of diversity is 1 - D, the probability that any two randomly
drawn species will be different. Values range from one (high diversity) to
zero (low diversity). It emphasizes common species and de-emphasizes rare
species, which means that the measure is not dramatically affected by missing
rare species during sampling. The effective number of species using the
Simpson's index is 1/D.
- Shannon-Wiener Index: This index is also very popular in ecological
studies. It is a measure of the “disorder” in a sample. The higher the disorder
or uncertainty, the more diverse a system is. The higher the H
0
value, the more
diverse the site. The Shannon-Wiener index is the negative sum of the
proportion of each species (p
i
) times the log of p
i
:
X
S
H
0
¼
p
i
ln
p
i
i
Index values typically range from 1.5 (low diversity) to 3.5 (high diversity).
This measure is more sensitive to rare species than Simpson's index, but less
sensitive than plain species richness. That is, rare species count for more
value in the Shannon-Wiener Index than in the Simpson. The effective
number of species using the Shannon-Wiener index is e
H
0
, the exponent of H
0
.
- Effective number of species: This metric can be calculated from any diver-
sity measure, and describes the equivalent number of equally common spe-
cies for a data set. That is, if all species were of equal abundance, how many
would there be? This number takes into account the evenness of the commu-
nity, and will always be lower than the actual species richness (unless all
species are equally abundant). Sites with higher numbers of effective species
are more diverse than sites with lower values. Unlike the index values, using
effective number of species makes intuitive sense to a lay audience.
- Evenness: An evenness value can also be calculated for each sample. Using
the Shannon-Wiener index (H
0
) as a starting point,
H
0
ln
S
J
¼
where S is the species richness. This metric is “Pielou's J,” and ranges from
one (perfect equitability among species) to zero (no equitability). In terms of
evaluating wetland plant communities, higher evenness may mean that a com-
munity is more diverse, and less dominated by a few highly competitive species.
5.4.3 Preparation of Multivariate Data
Users must be careful to understand the structure of their multivariate data before
beginning. Many parametric data analysis techniques rely on normally-distributed
