Transcript Soft” Approaches to Regional Species Pools for Plots

“Soft” Approaches to Regional
Species Pools for Plots
Tom Wentworth, Jason Fridley, Joel
Gramling, Todd Jobe
Ecoinformatics Working Group
November 25, 2002
What is a regional species pool?

Bob Ricklefs (TEON, 5e, 2001): “The
species that occur within a region are
referred to as its species pool. All the
members of the regional species pool
are potential members of each local
community.”
Local communities are subsets
of the regional species pool.

More from Bob Ricklefs (TEON, 5e,
2001): “A central concept of ecology is
that membership in local communities is
restricted to the species that can
coexist together in the same habitat.
Thus, each local community is a subset
of the regional species pool.”
Work of Weiher and Keddy…

Species sorting:
experimental study
of 20 wetland
species seeded into
120 wetland
microcosms
representing varied
environments
Bob Ricklefs (TEON, 5e, 2001): “Interactions of
species within local habitats make up only half
of the diversity equation.”
Regional vs. Local Effects
So what?

The relationship between the regional species
pool and local community is mediated by
important processes fundamental to our
understanding of how local communities are
organized:




dispersal
habitat selection
predatory and competitive exclusion
chance extinction
Interesting questions: (1) Is there
proportional sampling vs. saturation?
Interesting questions: (2) What is the
extent of nestedness?
We gain important insights from
examination of species pools.
Our Challenge: Building
Species Pools

We don’t know
the species pools
contributing to
our plots:


we could accept
arbitrary
definitions, but…
objective
approaches are
preferable: is
there a bottom-up
approach?
“Hard” vs. “Soft” Approaches
(sensu Fridley)

Hard: species are associated with one
another through co-occurrence in plots:


species pools are built through “chains” of cooccurrence among species
Soft: species pools are constructed as
plots/species are accumulated by “proximity”:



geographic (limited utility, but traditional)
environmental (attractive as we gather data)
compositional (most accessible)
Soft Pools: Geographic Basis

Place plots in a geographic space (x, y,
maybe z):



select a plot
accumulate species in the regional pool
from nearest neighbor plots
add species until…when???
Soft Pools: Geographic Basis

We don’t think this is necessarily the
best idea:



no well-defined stopping point
accumulating species through geographic
proximity builds pools with “strange
bedfellows” (consider the longleaf
savannah adjacent to a pocosin)…
but perhaps this is consistent with Ricklefs’
definition of regional species pools?
Soft Pools: Environmental Basis

Place plots in an environmental space



select a plot
accumulate species in the regional pool
from nearest neighbor plots
add species until you…


reach a plot that shares no species with
starting plot
reach some arbitrarily determined distance
Soft Pools: Environmental Basis

We like this idea:



support from work by Taylor, Aarssen et al.
builds pools using plots that are initially
similar from an environmental perspective
NCVS data base is richly endowed with
environmental data
Soft Pools: Compositional Basis

Place plots in an compositional space



select a plot
accumulate species in the regional pool
from nearest neighbor plots
add species until you…


reach a plot that shares no species with
starting plot
reach some arbitrarily determined distance
Soft Pools: Compositional Basis

We like this idea:


builds pools using plots that are initially
similar from a compositional perspective
not restricted by limited availability of
environmental data
Soft Pools: Alternatives

Plot-based environmental and
compositional spaces can also be
populated with species:
why not build pools based on species’
centers and accumulate these in a
nearest-neighbor approach?
 a nice start, but ignores differential
niche breadths of species…

Class data
ST N D 03
ST N D 09
ST N D 10
S T N D 0 1QUPH
SAAL
PITA
ST N D 17 ST N D 22
ULAL
ULRU
S
T N D 16
ST N D 11
PIEC
QUST
DIVI
Axis 2
LIST
JUVI
NYSY
COFL
ST N D 05
PRSE
CACA
ST N D 21
ST N D 20
FRPE
CATO
OXAR
ST N D 06
ULAM
LITU
ACRU
M ORU
AM AR CECA
ST N D 04
QUFA
QUM A
ST N D 07
QUVE
ST N D 23
ILOP
QUNI
CACO
S
BENI
T N D 26
ST N D 02
QUCO
ST
M
AVI
N D 08
QURU
M ATR
FAGR
OSVI
S T NID 1 2
QUM
QUAL
FRAM
CAGL
ST N D 18
ST N D 25
ACSA
ST N D 27
ST N D 24
Axis 1
ST N D 14
ST N D 15
Soft Pools: Alternatives

Plot-based environmental and
compositional spaces can also be
populated with species:

why not build pools based on distributions
of species overlapping a particular plot?
environmental or compositional gradient
Class data
ST N D 03
ST N D 09
ST N D 10
S T N D 0 1QUPH
SAAL
PITA
ST N D 17 ST N D 22
ULAL
ULRU
S
T N D 16
ST N D 11
PIEC
QUST
DIVI
Axis 2
LIST
JUVI
NYSY
COFL
ST N D 05
PRSE
CACA
ST N D 21
ST N D 20
FRPE
CATO
OXAR
ST N D 06
ULAM
LITU
ACRU
M ORU
AM AR CECA
ST N D 04
QUFA
QUM A
ST N D 07
QUVE
ST N D 23
ILOP
QUNI
CACO
S
BENI
T N D 26
ST N D 02
QUCO
ST
M
AVI
N D 08
QURU
M ATR
FAGR
OSVI
S T NID 1 2
QUM
QUAL
FRAM
CAGL
ST N D 18
ST N D 25
ACSA
ST N D 27
ST N D 24
Axis 1
ST N D 14
ST N D 15
Problems…

How many axes for environmental or
compositional space?



as number of axes increases, species pool
collapses to the species present in the plot
could limit analysis to n compositional or
complex environmental axes (from PCA),
but how many?
Edge effects limit detectability of
species pools for marginal plots