# Species diversity indices

Document Sample

```					Species diversity indices
Are these two communities equally diverse?

Site 1   Site 2
Sp A   12       49
Sp B   11       1
Sp C   14       1
Sp D   13       1
3 Ways to include abundance in
species diversity

1. Plot the distribution of individuals
amongst species.

2. Summarize both abundance and species
richness in a single index.

3. Examine the evenness of the distribution
of individuals amongst species
1. Plot the distribution of abundance
amongst species
Three models have been proposed:
• Log-series
• Broken stick
• Log-normal

All are supported by data

None have a good theoretical explanation
(save Stephen Hubbell’s neutral model, 2001)

Hard to tell apart statistically!
Log-series

Most species are represented by only a couple
of individuals (i.e. rare). Only a few highly-
abundant (i.e. common) species.
Log abundance

100
Straight line predicted
per species

10

1

1   2   3

Linear species rank (1=most abundant)
Broken-stick
• More equitable distribution of abundances amongst
species than log-series.

• Like log series most species have a fairly low
abundances.
abundance per

50

40
Straight line predicted
species

30
Linear

20

10

1   10    100

Log Species rank (1=most abundant)
Log-series revisited
“Most species are rare”

Number of species
Log abundance

100
per species

10

1

1   2   3
Log abundance class per
Linear species rank                             species
(1=most abundant)
Log-normal
Most species are do NOT have abundances of only a
few individuals, but rather have intermediate
abundances (on a log scale! Still low)
Number of species

Log normal

Log abundance class per species
2003 class mite data
Essentially number of species

Log2 abundance class

PRIMER “Geometric class plot”
2003 class mite data
Linear abundance per species

Log species rank

PRIMER “Dominance plot”
2. Summarize everything in one index
Some indices output by PRIMER (formulas in Krebs
and Magurran)

• Simpson’s (1-lambda, or 1-D in Krebs)

• Shannon-Wiener

• Alpha (a parameter from log series)

• Margalef d
Shannon-Wiener =
sum (-Proportion spA * ln (prop spA)+
(-Proportion spB*ln(prop spB)...)

Site 1   Site 2    Site 1 Site 2
-p*lnp -p*lnp
Sp A     12       49          0.34    0.06
Sp B     11       1           0.33    0.08
Sp C     14       1           0.36    0.08
Sp D     13       1           0.35    0.08
sum                           1.38    0.28
Which index?

Read Krebs and Magurran and consider:

- Sensitivity to differences in sample size.

- Do you want differences in rare or abundant
species to be emphasized?

- Do you want differences in species richness or
evenness to be emphasized?

- How does log-normal vs. log-series affect?

- Performance in other studies (what works?).
3. Measure evenness
separately

Pielou’s J: Comparison of actual Shannon-Wiener
with Shannon-Wiener if species had equal
proportion (log S).

•Close to 1: very even distribution of abundances
amongst species

•Close to 0: very uneven
Site 1   Site 2   Site 1 Site 2
-p*lnp -p*lnp
Sp A 12         49          0.34    0.06
Sp B 11         1           0.33    0.08
Sp C 14         1           0.36    0.08
Sp D 13         1           0.35    0.08
Sum (Shannon-Wiener)        1.38    0.28
Log S                       1.39    1.39
J’                          1.00    0.20
Smarties revisited!

1. Count your smarties. Each colour is a different
species.

2. Does your distribution fit log-series or log-normal?
(Use log 2 abundance classes to make bar graph).

3. Calculate the following diversity/evenness indices:

 Simpson’s
 Shannon-Wiener
 Pielou’s J

```
DOCUMENT INFO
Shared By:
Categories:
Stats:
 views: 37 posted: 4/14/2011 language: English pages: 20