Transcript Niche Overlap

Niches
• What is a niche?
Niche Theory
• You can think of it as its ‘address’
What is a niche?
• A multidimensional
expression
of where or
how a
species
lives
Niche Theory
• Generalist vs. Specialist
Niche Theory
• Why the hump in the curve?
Because there is an optimal
‘size’ or space on the resource
continuum
• Consequently, they taper off at each end
of the resource spectrum becoming less
competitive at either end
Realized vs. Fundamental
• Why might a sp. not completely occupy its
fundamental niche?
Fundamental vs. Realized
Niche Breadth
• What are some of the factors that
impact niche breadth?
• Competition
• Predation
• Resource predictability
• Resource abundance
• Intraspecific competition (ind. vs. pop)
Niche Theory
• Despite its conceptual simplicity, observing
niche competition is not so easy
• A basic premise, in the presence of a
strong competitor, niche breadth should
change (short term through behavioral
modifications and long term through
evolutionary adaptations)
Fundamental vs. Realized
Niche Theory
• Huey et al. (1974) demonstrated that 2 sp.
of skinks appear to have large dietary
overlap of termites when population
adjacent, but shift away from one another
when sympatric
• Furthermore, at the end of sympatric
distribution, the smaller sp. quickly
displayed a significant increase in body,
head, and jaw size
Niche Theory
• This dramatic change is frequently
associated with character displacement
• What is the problem with documenting
character displacement?
Niche Theory
• We may observe current patterns in an
attempt to determine past events
• One example is that adaptive radiations
and subsequent ‘regular spacing’ on
resource axes
Niche Theory
• However, different size bills and
bodies can also simply reflect diet
specialization (i.e. just getting better)
• There are several clusters of similar
species that have developed
differences in foraging areas and that
is an indirect indication that
competition may have been at work
(previously)
Niche Theory
• Tyrannid flycatchers; perhaps an
example of past competition
Niche Theory
• Foraging
relationship
among several
antbirds
(Formicariidae)
showing
ecological
separation
Niche Theory
• Projected niche
relationships in 2
resource space
Niche Theory
• Ecological
segregation should
work to minimize
competition and
niche overlap
What is niche overlap?
• Species can be
generalists and
have relatively high
niche overlap and
compete strongly
with other species
What is niche overlap?
• Species can either
separate and
minimize
competition (lower
the niche overlap)
by becoming a
resource specialist
(a)
Niche Breadth (width or size)
• Some plants and animals are more
specialized than others, and measures of
niche breadth attempt to quantify this
• It is typically measured by observing the
distribution of individual organisms within a
set of resource states
• Information is collected and presented in
a resource matrix
Example of a Resource Matrix
The percentage utilization of 14 microhabitats by 11 species of SW desert lizards
What are common resource
states?
• Resource states may be defined in a
variety of ways:
• 1) Food Resources: taxonomic identity of
food taken may be used as a resource
state, or the size category of food item
(without regard to taxonomy) could be
defined as the resource state
What are common resource
states?
• 2) Habitat Resources: habitats for animals may
be defined botanically or from physical-chemical
data in a series of resource states
• 3) Natural Sampling Units: sampling units like
lakes or leaves or individual fruits may be
defined as resource states
• 4) Artificial Sampling Units: a set of random
quadrats may be considered different resource
states
Niche Overlap
• Shared niche space
of the Hutchinsonian
multidimensional
niche
• Thought to represent
a measure of
competition intensity
Levin’s Measure
• Niche breadth be estimated by measuring
the uniformity of distribution of individuals
among the resource states
B = Y2 / ∑ N2j
where B = Levins measure of niche breadth
Nj = Number of individuals found in or
using resource state j
Y = ∑ Nj = Total no. of individuals
sampled
Shannon-Wiener Measure
• Others have suggested using the S-W
formula from information theory to
measure niche breadth
H’ = -∑ pj log pj
• This measure tends to weight rare
resources (compared to Levin’s, which
weights common resources)
Niche and Communities
• Community change in the limiting similarity
model comes about through repeated
colonization and extinction of species with
different utilization curves
• If adjacent species are ‘too close’ together,
one of the pair will go extinct, depending
on the overlap and the carrying capacity of
the environment (Gause’s competitive
exclusion principle)
Niche and Communities
• After repeated C & E events, an equilibrium is
reached with a maximum number of coexisting
species separated
by a critical
minimum spacing
(Gotelli Fig. 4.1)
Niche Overlap
• One way to understand community
organization is to measure the overlap in
resource use among the different species
in a community
• The most common resources measured in
order to calculate overlap are food and
space
• Several measures have been proposed,
with various strengths and weaknesses
Niche Overlap
• Historically, analyses of niche overlap were
based on the theory of similarity (MacArthur and
Levins 1967)
• However, this early measure was asymmetrical
and has since been replaced by a more
symmetrical measure
• Overlaps calculated this way have been equated
with the competition coefficients of the LotkaVolterra equations and are thus proportional to
the intensity of competition
MacArthur and Levins Overlap
∑ p2i p1i
O21 =
∑ (p1i)2
Asymmetrical Competition
(impact of sp1 on sp2 is not the
same as that of sp2 on sp1)
Pianka’s Overlap
∑ p2i p1i
O12 = O21 =
√ ∑ (p2i)2 ∑ (p1i)2
Czekanowski’s Index
O12 = O21 = 1 – 0.5 ∑ |p1i – p2i |
Percentage Overlap
• This is a very attractive measure as it is
relatively easy to calculate and interpret
Pjk = [ ∑ (minimum pij, pik)] * 100
where Pjk = percent overlap between j and k
pij & pik = proportions resource i is of the
total resources used by species j and k
and n = total number of resource states
Morisita’s Measure
C = 2 ∑ pij * pik / ∑n pij[(nij-1)/(Nj-1)] +
∑n pik [(nik-1)/(Nk-1)]
This measure is free from bias over a range
of possible values and is a relatively good
measure of overlap
Null Model Decisions
• Weighted vs. Unweighted Indices
• If all resources states are not equally
available, observed overlaps in utilization
may not accurately reflect similarity in use
• In particular, if some resource states are
extremely common and others are
extremely rare, species may appear very
similar in their resource utilization
“electivity”
• Ecologists have suggested modifying
existing indices to account for the
electivity, (the relative ability (or
preference), of resource use
• Incorporating resource availability may
have a major effect on measures of
overlap (e.g. think about large use; could
be abundant resource or high preference)
Pairwise Niche Overlap
EcoSim
Weighted vs. Unweighted
• If resource states are not equally abundant,
observed utilizations will tend to overestimate
the amount of ecological overlap (i.e. everyone
is using abundant resources)
• However, it can also correct for uneven
resources
• For example, only 2 of 10 mean utilization
overlaps for Pianka’s (1967) NA lizard
communities differed from null models whereas
all 10 mean electivities differed significantly
• This approach has problems as well, see book
• To generate a null model to test for
deviations from expected (for overlaps),
we need to construct a null model
• We could 1) randomize the dietary or
activity data OR 2) randomize species
occurrences
Niche Overlap &
Species Occurrences
• If competition limits niche overlap, then the
particular combination of species that
coexist on the island should have lower
overlap than a randomly assembled set of
species from the same source pool
(Gotellli and Graves)
Randomization of species
occurrences
• How to generate a null model for niche overlap
of 18 species of lizards on 37 islands in the Sea
of Cortez?
• Using biologically realistic criteria, Case (1983)
identified a source pool of 18 mainland species
that could potentially colonize each island
• For each island with i species, Case enumerated
all the unique (1i8) combinations of exactly i
species as null communities
Example
Species coexisting on islands had lower niche overlap
(30 of 37 times) than would be expected in the
absence of competition (or a nonrandom pattern of
resource availability)
• however, this analysis assumed that
species colonized islands equiprobably
• When %occupied was used instead, only
23 of 27 fell below the median
• This suggests that dispersal ability may
have contributed to the pattern of reduced
overlap
Wait…
• What is low overlap was a result of
nonrandom patterns of resource
availability on islands?
• If the same nonoverlapping sets of
resources were present on several
islands, the same combinations of lowoverlap species would be found
• For most island size classes, an
improbably small number of species
combinations was represented
• SO? This suggests the same low-overlap
configurations tended to recur.
• Consequently, the pattern initially
observed probably resulted from a
nonrandom distribution of resources and
NOT competition
Null Model Example
• Schoener (1988) also examined niche overlap of
island lizard species sampled from a larger
source pool, but examined microhabitat use
• Coexisting species usually differed in the
structural habitats they occupied
• e.g. on two-species islands, each species
occupied a different category; coexistence in the
same habitat was found once on 3-species
islands and never on 4-species islands
Null Model Example
• He tested four different ‘source pool’
scenarios, varying in the likelihood of
occupying habitat categories
• Together (Schoener and Case), these
studies show results will be sensitive to
sample size, source pool definitions, and
assumptions about the colonization
potential of species…and is a good tool for
evaluating niche overlap
Randomization of Utilization
Matrices
• In most cases, a source pool is not
available to generate a biologically realistic
community
• Instead, the observed utilization matrix
must be used to estimate overlap values in
the absence of competition
Lawlor’s Algorithms
• Four algorithms that vary in the amount of
original utilization data is retained in the
null community
0 states
• Obs. utilization from uniform
distribution
• Obs. utilizations reshuffled
0 states
randomized retained
RA1
RA2
RA 3
RA4
Lawlor’s Algorithms
• RA1: all resource state is possible and
equiprobable
• RA2: resource utilization is randomized,
but only for those states that are >0
• RA3: observed utilizations are randomly
reassigned to different resource categories
• RA4: only the nonzero resource states in
each row are reshuffled
Lawlor’s Algorithms
• RA1: all resource state is possible and
equiprobable
• RA2: resource utilization is randomized,
but only for those states that are >0
• RA3: observed utilizations are randomly
reassigned to different resource categories
• RA4: only the nonzero resource states in
each row are reshuffled
Lawlor’s Algorithms
• RA1: all resource state is possible and
equiprobable
• RA2: resource utilization is randomized,
but only for those states that are >0
• RA3: observed utilizations are randomly
reassigned to different resource categories
• RA4: only the nonzero resource states in
each row are reshuffled
Lawlor’s Algorithms
• RA1: all resource state is possible and
equiprobable
• RA2: resource utilization is randomized,
but only for those states that are >0
• RA3: observed utilizations are randomly
reassigned to different resource categories
• RA4: only the nonzero resource states in
each row are reshuffled
RA3: random reassignment
Hab A
Sp A 0.3
Sp B 0.7
Sp C 0.1
Hab B
0.5
0.0
0.2
Hab C
0.0
0.2
0.5
Hab D
0.2
0.1
0.2
Hab A
Sp A 0.5
Sp B 0.1
Sp C 0.5
Hab B
0.2
0.7
0.2
Hab C
0.3
0.0
0.2
Hab D
0.0
0.2
0.1
Lawlor’s Algorithms
• RA1: all resource state is possible and
equiprobable
• RA2: resource utilization is randomized,
but only for those states that are >0
• RA3: observed utilizations are randomly
reassigned to different resource categories
• RA4: only the nonzero resource states in
each row are reshuffled
Sp A
Sp B
Sp C
Sp A
Sp B
Sp C
Hab A
0.3
0.7
0.1
Hab A
0.2
0.2
0.2
Hab B
0.5
0.0
0.2
Hab B
0.3
0.0
0.5
Hab C
0.0
0.2
0.5
Hab C
0.0
0.7
0.2
Hab D
0.2
0.1
0.2
Hab D
0.4
0.1
0.1
Which one?
• RA1 is consistent with the idea that
competition is so severe that some
species are completely denied the use of
certain resources by the presence of
competitors
• Conceptually, resource is in its theoretical
niche, just not its current realized niche
(not recommended by Gotelli and Graves)
Which one?
• RA2 & RA4 (not recommended) ensure
that species which do not use certain
resource states in nature never do so in a
null community
• RA3 is a compromise by retaining the
same number of zero states as the
originally observed, but does not constrain
those zeros to their original placement