photo of s. p. hubbell from ucla; quotes from hubbell (2001, pg. 30)
DESCRIPTION
Relative-Abundance Patterns. “ No other general attribute of ecological communities besides species richness has commanded more theoretical and empirical attention than relative species abundance ”. - PowerPoint PPT PresentationTRANSCRIPT
“No other general attribute of ecological communities besides species richness has commanded more
theoretical and empirical attention than relative species abundance”
“Commonness, and especially rarity, have long fascinated ecologists…, and species abundance is of central theoretical and practical
importance in conservation biology…”
“In particular, understanding the causes and consequences of rarity is a problem of profound significance because most species are uncommon
to rare, and rare species are generally at greater risk to extinction”
Photo of S. P. Hubbell from UCLA; quotes from Hubbell (2001, pg. 30)
Relative-Abundance Patterns
Empirical distributions of relative abundance (two graphical representations)
Data from BCI 50-ha Forest Dynamics Plot, Panama; 229,069 individual trees of 300 species; most common species nHYBAPR=36,081; several rarest species nRARE=1
1. Frequency histogram(Preston plot)
0
5
10
15
20
25
30
35
40
45
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Log2 abundance
Fre
quen
cy (
no.
of s
pp.)
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 50 100 150 200 250 300 350
2. Dominance-diversity (or rank-abundance)
diagram, after Whittaker (1975)
Rank
Log 1
0 a
bund
ance
Relative-Abundance Patterns
1. Descriptive (inductive):
Fitting curves to empirical distributions
2. Mechanistic (deductive):
Creating mechanisticmodels that predict theshapes of distributions
Approaches used to better understand relative abundance distributions
A combined approach: create simple, mechanistic models that predictrelative abundance distributions; adjust the model to maximize
goodness-of-fit of predicted distributions to empirical distributions
Relative-Abundance Patterns
Preston (1948) advocated log-normal distributions
Num
ber
of s
peci
es
1 16 256 4096
60
40
20
Log normal fit to the data
Figure from Magurran (1988)
Log2 moth abundance (doubling classes or “octaves”)
The left-hand portion of the curve may be missing beyond the “veil line” of a small sample
1 16
20
Preston (1948) advocated log-normal distributions
Figure from Magurran (1988)
Num
ber
of s
peci
es
Log2 moth abundance (doubling classes or “octaves”)
1 16 256
40
20
The left-hand portion of the curve may be missing beyond the “veil line” of a small sample
Preston (1948) advocated log-normal distributions
Figure from Magurran (1988)
Num
ber
of s
peci
es
Log2 moth abundance (doubling classes or “octaves”)
1 16 256 4096
60
40
20
Log-normal fit to the data
The left-hand portion of the curve may be missing beyond the “veil line” of a small sample
Preston (1948) advocated log-normal distributions
Figure from Magurran (1988)
Num
ber
of s
peci
es
Log2 moth abundance (doubling classes or “octaves”)
Figure from Magurran (1988)
Rank
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)Variation among hypothetical rank-abundance curves
Figure from Magurran (1988)
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)
Rank
Examples of relative-abundance data; each data set is best fit by one of the theoretical distributions
geometric series
broken-stick
log-normal
Log-normal distribution:
Log of species abundances are normally distributed
Why might this be so?
Central Limit Theorem: When a large number of factors combine to determine the value of a variable (number of individuals per species), random variation in each of those factors (e.g., competition, predation, etc.) will result in the variable being normally distributed
May (1975) suggested that it arises from the statistical properties of large numbers and the Central Limit Theorem
J. H. Brown (1995, Macroecology, pg. 79):“…just as normal distributions are produced by additive combinations of random variables, lognormal distributions are produced by multiplicative combinations of random variables (May 1975)”
Figure from Magurran (1988)
Rank
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)
Figure from Magurran (1988)
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)
Rank
geometric series
broken-stick
log-normal
Geometric series:
The most abundant species usurps proportion k of all available resources, the second most abundant species usurps proportion k of the left-overs, and so on down the ranking to species S
…
Sp. 1
Sp. 2
Sp. 3
Sp. 4
Total Available Resources
Does this seem like a reasonable mechanism, or just a metaphor with uncertain relevance to the real world?
Figure from Magurran (1988)
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)
Rank
geometric series
broken-stick
log-normal
Geometric series:
This pattern of species abundance is found primarily in species-poor (harsh) environments or in early stages of succession
Remember, however, that pattern-fitting alone does not necessarily mean that the mechanistic metaphor is correct!
Figure from Magurran (1988)
Rank
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)
Figure from Magurran (1988)
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)
Rank
broken-stick
log-normal
geometric series
Broken Stick:
The sub-division of niche space among species may be analogous to randomly breaking a stick into S pieces (MacArthur 1957)
Sp. 1
Sp. 2
Sp. 3
Sp. 4
Total Available Resources
Sp. 5
This results in a somewhat more even distribution of abundances among species than the other models, which suggests that it should occur when an important resource is shared more or less equitably among species
Even so, there are not many examples of communities with species abundance fitting this model
Figure from Magurran (1988)
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)
Rank
broken-stick
log-normal
geometric series
Broken stick:
Uncommon pattern
In any case, remember that pattern-fitting alone does not necessarily mean that the mechanistic metaphor is correct!
Figure from Magurran (1988)
Rank
Abu
ndan
ce (
% o
f tot
al in
divi
dual
s)
Num
ber
of s
peci
es
1 16 256 4096
60
40
20
Figure from Magurran (1988)
Log2 moth abundance (doubling classes or “octaves”)
Log series fit to the data
Note the emphasis on rare species!
Log normal fit to the data
Log series:
First described mathematically by Fisher et al. (1943)
Log series takes the form: x, x2/2, x3/3,... xn/n
where x is the number of species predicted to have 1 individual, x2 to have 2 individuals, etc...
Fisher’s alpha diversity index () is usually not biased by sample size and often adequately discriminates differences in diversity among communities even when underlying species abundances do not exactly follow a log series
You only need S and N to calculate it… S = * ln(1 + N/)
But do we understand why the distribution of relative abundances often takes on this form?
Photo of R. A. Fisher (1890-1962) from Wikimedia Commons
Hubbell’s Neutral Theory
An attempt to predict relative-abundance distributions from neutral models of birth, death, immigration, extinction, and speciation
Assumptions: individuals play a zero-sum game within a communityand have equivalent per capita demographic rates
Immigration from a source pool occurs at random; otherwise species composition in the community is governed by community drift
Neutral model of localcommunity dynamics
Figure from Hubbell (2001)
Hubbell’s Neutral Theory
Neutral model of localcommunity dynamics
Figure from Hubbell (2001)
Hubbell’s Neutral Theory
Neutral model of localcommunity dynamics
Figure from Hubbell (2001)
Hubbell’s Neutral Theory
The source pool (metacommunity) relative abundance isgoverned by its size, speciation and extinction rates
Local community relative abundance is additionallygoverned by the immigration rate
By adjusting these (often unmeasurable) parameters, one is able to fit predicted relative abundance distributions to those observed in empirical
datasets
Hubbell’s Neutral Theory
An attempt to predict relative-abundance distributions from neutral models of birth, death, immigration, extinction, and speciation
Assumptions: individuals play a zero-sum game within a communityand have equivalent per capita demographic rates
Immigration from a source pool occurs at random; otherwise species composition in the community is governed by community drift
Theta () = 2JM
JM = metacommunity size
= speciation rate
The model provides predictions that match
nearly all observed distributions of relative
abundance
Figure from Hubbell (2001)
Hubbell’s Neutral Theory
Theta () = 2JM
JM = metacommunity size
= speciation rate
m = immigration rate
Model predictions can be fit to the observed relative-abundance
distribution of trees on BCI
Figure from Hubbell (2001)
Hubbell’s Neutral Theory
Theta () = 2JM
JM = metacommunity size
= speciation rate
m = immigration rate
Does the good fit of the model to real data mean that we now understand
relative abundance distributions
mechanistically?
Figure from Hubbell (2001)
Hubbell’s Neutral Theory
Low Plateau 24.80 ha High Plateau 6.80 ha
Swamp 1.20 haStream 1.28 ha
Slope 11.36 ha
Mixed 2.64 ha Young 1.92 ha
7 topographically / hydrologically / successionallydefined habitats
Barro Colorado Island 50-ha plot, Panama
0
1
2
3
4
5
0 50 100 150 200
Relative abundance rank
Lo
g10
(N)
Positiveassociationwith swamp
Neutralassociationwith swamp
Negativeassociationwith swamp
171 species,each with> 60 stems
3355 total stems in the 1.2-ha swamp
Relative abundance is often a function of “Biology”
0
1
2
3
4
5
0 50 100 150 200 250 300 350
Relative abundance rank
Lo
g10
(N) Ficus &
Cecropia
Non-Moraceae
OtherMoraceae
303 species,each with atleast 1 stemIn 1990
Relative abundance is often a function of “Biology”