Dynamics of Regional Distribution: The Core and Satellite Species Hypothesis
FOM 30 - Hanski 1982: Dynamics of Regional Distribution:
The Core and Satellite Species Hypothesis
The Core and Satellite Species Hypothesis
Author or coauthor of 200+ publications, elected to the National Academy of Sciences & the
Royal Society.
Royal Society.
Dr. Metapopulation to many scientists
bachelor's degree and Licentiate's degree: University of Helsinki, 1976
Doctoral degree: University of Oxford,1976 - 1979.
Most widely known for his seminal contributions to metapopulation ecology, both theoretical
and empirical. But he also made many important and wide-ranging contributions to other
arenas of ecological inquiry, including in particular predator—prey, host—parasitoid, and
host—pathogen interactions (from Holt 2017. Ilkka Hanski, The “Compleat Ecologist”:
An Homage to His Contributions to the Spatial Dimension of Food Web Interactions.
Where he intended to “pay homage to the many contributions that Ilkka Hanski made to our
understanding of the vertical structure of ecological communities”.)
and empirical. But he also made many important and wide-ranging contributions to other
arenas of ecological inquiry, including in particular predator—prey, host—parasitoid, and
host—pathogen interactions (from Holt 2017. Ilkka Hanski, The “Compleat Ecologist”:
An Homage to His Contributions to the Spatial Dimension of Food Web Interactions.
Where he intended to “pay homage to the many contributions that Ilkka Hanski made to our
understanding of the vertical structure of ecological communities”.)
Assistant professor in the School of Biological Sciences at the U of Nebraska-Lincoln
Department of Paleobiology, National Museum of Natural History - Evolution of Terrestrial Ecosystems
Curator
Curator
Postdoc at UC Santa Barbara, Nat’l Center for Ecological Analysis and Synthesis
Postdoc at UNM - with F. A. Smith and J. H. Brown
Ph.D. University of Chicago, 2001
Studies changes in faunal diversity in the fossil record, uses particularly the mammal fossil
record to evaluate how current changes in global climate may affect diversity patterns in the future
and how the effects of global climate change can affect species diversity. “Because it provides a
useful way to compare modern species and communities to fossil species and communities,” she
“focuses on the similarities and differences in macroecological patterns across space and time.”
record to evaluate how current changes in global climate may affect diversity patterns in the future
and how the effects of global climate change can affect species diversity. “Because it provides a
useful way to compare modern species and communities to fossil species and communities,” she
“focuses on the similarities and differences in macroecological patterns across space and time.”
1. Introduction
“…An understanding of both spatial processes (distribution of the species in physical space)
and resource partitioning (distribution of the species in niche space) are essential components
to a satisfactory explanation of the perennial questions: Why are there so many species?
Why are there so many rare species?”
and resource partitioning (distribution of the species in niche space) are essential components
to a satisfactory explanation of the perennial questions: Why are there so many species?
Why are there so many rare species?”
Definitions
Abundance: number of individuals at a local population site
Distribution: number of population sites occupied by the species
2. Local abundance and regional distribution are independent
“...are local abundance and regional distribution independent of each other? The answer is no.”
This is a revision to Levins’ model. The positive correlation between species abundance and
regional distribution conclusion is backed up with data on the abundance and distribution of
four species of invertebrates from a variety of biomes (e.g. Figure 1)
regional distribution conclusion is backed up with data on the abundance and distribution of
four species of invertebrates from a variety of biomes (e.g. Figure 1)
3. Models of distribution
He uses the Lotka logistic equation to describe dynamics in local abundances (eq 1) and the
general model proposed by Levins to model distributions (eq 2). He then uses the stochastic
version of eq 2, also from Levins, assuming that the extinction parameter (e) is a random variable,
and be able to analyse the distribution of p as phi(p) for either a single species in a long period of
time or for a community of similar species at a given moment
general model proposed by Levins to model distributions (eq 2). He then uses the stochastic
version of eq 2, also from Levins, assuming that the extinction parameter (e) is a random variable,
and be able to analyse the distribution of p as phi(p) for either a single species in a long period of
time or for a community of similar species at a given moment
If i (colonization rate) does not depend on p, then the population exhibits logistic growth.
phi(p) is bimodal if the variance is greater than s (the difference between i - e’).
There is one stable equilibrium point → ^p = 1 If the difference between colonization rate and
extinction rate is smaller than the variance divided by 3, then the distribution of p is bimodal. So,
instead of e being purely stochastic, Hanski alters the equation so that e decreases with p. e goes
down with increasing population size and up with decreasing population size. e is more sensitive to
small sizes and that there is a point at which increased N doesn’t increase e.
extinction rate is smaller than the variance divided by 3, then the distribution of p is bimodal. So,
instead of e being purely stochastic, Hanski alters the equation so that e decreases with p. e goes
down with increasing population size and up with decreasing population size. e is more sensitive to
small sizes and that there is a point at which increased N doesn’t increase e.
4. Ecological appraisal
Hanski’s modification of Levins’ model changes its conclusion: Levins arrived at the conclusion that
populations would “hover around a stable internal equilibrium,” while Hanski’s modifications show that,
with sufficient stochasticity within the model, populations are either moving towards p=0 or p=1,
extinction or maximal distribution. Hanski’s model is supported by data from several other studies,
which show a bimodal distribution of species’ regional distributions (e.g. fig. 12, Hanski unpubl.).
populations would “hover around a stable internal equilibrium,” while Hanski’s modifications show that,
with sufficient stochasticity within the model, populations are either moving towards p=0 or p=1,
extinction or maximal distribution. Hanski’s model is supported by data from several other studies,
which show a bimodal distribution of species’ regional distributions (e.g. fig. 12, Hanski unpubl.).
phi(p) gives the distribution of p both for a single species during a long period of time, and
for a community of similar species at a given moment. To test the latter qualitatively, the
model “requires that all the species may establish local populations at the same sites, and
that interspecific influences on model parameters are density- and frequency-independent.
[Hanski’s] model then predicts bimodality of p's, peaks close to unity and zero, while Levins's
model predicts unimodality with the peak not very close to unity or zero."
for a community of similar species at a given moment. To test the latter qualitatively, the
model “requires that all the species may establish local populations at the same sites, and
that interspecific influences on model parameters are density- and frequency-independent.
[Hanski’s] model then predicts bimodality of p's, peaks close to unity and zero, while Levins's
model predicts unimodality with the peak not very close to unity or zero."
5. More about testing the core-satellite hypothesis
Core and satellite species have distinct mechanisms underlying their population dynamics, yet their
population dynamics are linked. A satellite species may become a core species, or go extinct. A core
species may go extinct, but along the way it will become a satellite species. Under this model, these
transitions may occur even in a constant environment. Under the species-area hypothesis (number
of species increases with area), the number of satellite species should decrease as area decreases.
population dynamics are linked. A satellite species may become a core species, or go extinct. A core
species may go extinct, but along the way it will become a satellite species. Under this model, these
transitions may occur even in a constant environment. Under the species-area hypothesis (number
of species increases with area), the number of satellite species should decrease as area decreases.
6. The core-satellite hypothesis and r-K theory
The r-K theory and the core-satellite hypothesis are both based on the logistic equation. Core and
satellite species are determined by “stochastic variation in spatial population dynamics.”
satellite species are determined by “stochastic variation in spatial population dynamics.”
“Although the two concepts are fundamentally different, core species are related to K-species, and,
less obviously, satellite species are related to r-species.”
less obviously, satellite species are related to r-species.”
7. A historical perspective
Frequency vs. distribution (between-site occurrence)
The bimodal frequency of distribution has been found before by Raunkiaer in the early 1900s, but the
idea gradually fell out of favor by the 1960s.
idea gradually fell out of favor by the 1960s.
8. Concluding remarks - visiting Hutchinson’s niche space
Why is the core-satellite hypothesis needed? It is assumed that competition will result in more even
distribution in niche space. But what is the null hypothesis for this?
distribution in niche space. But what is the null hypothesis for this?
Niche spacing under the core-satellite hypothesis should result in more even spacing in core
populations, where organisms are interacting more frequently and for longer periods of time.
populations, where organisms are interacting more frequently and for longer periods of time.
“Hence, core species are not expected to be a random sub-set of all the species with respect to
niche position; we expect core species to comprise such a sub- set within which species are better
spaced-out from each other than species are within a truly random sub-set.”
niche position; we expect core species to comprise such a sub- set within which species are better
spaced-out from each other than species are within a truly random sub-set.”
Discussion Questions:
1) What parallels do you see between the core-satellite hypothesis and Rabinowitz’s seven
forms of rarity?
forms of rarity?
2) Hanski says that this model can be used in two ways, with a single species through time
or with several closely related species in a community. Are you convinced that this model
could be used with closely related species in a community?
or with several closely related species in a community. Are you convinced that this model
could be used with closely related species in a community?
Comments
2. I thought that this paper was helping to provide evidence that this model could be used with closely related species in a community, that you have species that tend to go to their maxima (1) or tend towards extinction (0) that created this core-satellite phenomenon with groups of species. I could see this also being true for a single species over long periods of time, but would like to see a similar analysis to provide evidence towards that.
2) I think this method is appropriately used when applied to multiple species in a community, as shown here by Hanski. It makes sense that local abundance of a species depends on abundance of closely related species, because of species interactions.
Another interesting tie in this paper to other hypotheses was relating the relationship between species diversity and area discussed here to island biogeography theory. It makes sense that as islands get smaller, there are less satellite species because overall species diversity decreases and the satellite species would be the first to disappear because they are locally rare.
2) I think Hanski argues that this model works best for comparing similar species as a a disclaimer. Basically, it wouldn't work the same if we compare an organism that might be generalized vs those that are highly specialized (water striders specific to particular lakes was used as one example).
This model could work for species that are closely related but I like Mariah’s point of considering the highly specialized species that exist in a really distinct niche space and range.
One aspect of this paper I really enjoyed in our zoom discussion was the species range shifting from satellite to core, and what are the parameters that go into those changes. When looking at ranges, and their destruction, interruption or overall evolution (human caused or not), I would be really interested to see what exactly causes these shifts over time and how that could be introduced into this model.