Transcript Lecture_wk5_Suding

Trajectories of Community Change:
using traits to understand
convergence and divergence
When are traits important?
“the agony of community ecology”
• Convergence and Ecological Dynamics
– Deterministic, Contingent, Stochastic
• Mechanisms
– Local interactions
– Regional processes
– Productivity and disturbance
• Few experimental tests
– Even fewer that link convergence with traits
• Synthesis and hypotheses to test
Determinism
Convergence to a
stable
deterministic
endpoint
Same conditions =
same structure
Getis, Getis, Feldman, Introduction to Geography, 11th ed.,
2008
Global Convergence
Photos © Encyclopedia Britannica
Predictable and parallel changes in the trait values
across a productivity gradient
Diaz et al. 2004 JVS
Wright et al. 2004
Transition from the acquisitive (‘fast and leaky’) to conservative (‘slow
and tight’) strategies
Field patterns indicate relatively large role of
contingency
Environmental factors only explained
10% of variability among communities
McCune and Allen 1985 Can J Botany
Narrowing the scale:
Community Assembly
• Different components of community structure change
differently
• Different views of community dynamics
• When should assembly be deterministic?
Views of Community Assembly
• Deterministic: environmental conditions and interspecific
interactions cause community structure to converge.
Shugart et al. 1988
Trowbridge 2007, EcolAp
Views of Community Assembly
• Deterministic: environmental conditions and interspecific
interactions cause community structure to converge.
• Contingent: Stochasticity, priority effects, and random drift
causes community structure to diverge (or not to
converge).
– Both stochastic or deterministic (alternative state) dynamics
diverge, but differ in the number and predictability of endpoints.
Cramer and Hobbs, 2006
Scheffer, 2004
Contingent effects can be deterministic, but still
cause divergence. Are traits less predictive?
Single Equilibrium
Threshold
Alternative States
Cyclic
Mechanisms
• Local interactions:
– negative feedbacks (resource partitioning, negative freq
dependence) lead to high local diversity and convergence
– Positive feedbacks (priority effects, self-beneficial species effects
on ecosystem processes) lead to high regional diversity and
divergence
Sp A
Sp B
Sp C
Time
Dissimilarity among plots
Relative Abundance
• E.g., for positive feedbacks:
Time
Mechanisms
• Local interactions:
– negative feedbacks (resource partitioning, negative freq
dependence) lead to high local diversity and convergence
– Positive feedbacks (priority effects, self-beneficial species effects
on ecosystem processes) lead to high regional diversity and
divergence
Sp A
Sp B
Sp C
Time
Dissimilarity among plots
Relative Abundance
• E.g., for negative feedbacks:
Time
While evidence for both negative and positive
feedbacks exist, less clear how they relate to traits
Scaled Population Growth Rate/Yr
Measure different
types of traits? E.g.,
ones that indicate
complementarity at
the community
level.
Initial Frequency
Harpole and Suding 2007
Mechanisms
• Regional processes:
–
–
–
–
Dispersal limitation lead to divergence due to chance colonization
Large regional pool lead to divergence (Law and Morton 1993)
Strong connectedness lead to convergence
Distance as a proxy for dispersal
Chase (2003)
Field patterns: dispersal
limitation
Dispersal limitation (distance as proxy): study plots
that were far apart became more similar over time.
Environmental filtering (elevation as proxy): study
plots with a large elevation differences became less
similar over time
But then processes REVERSED direction in
southern floodplain.
Trowbridge 2007 Ecol App.
Dispersal limitation?:
species-level priority effects and
trait-based assembly rules
Composition was compared using
Euclidian distance in the first four PCA
composition axes
Trait groups constructed using hierarchical
clustering (Ward’s method) of 87 species
and 17 traits into 14 groups.
Contrast indicative of dispersal limitation?
Fukami et al, 2005
Dispersal limitation?:
species-level priority effects and
trait-based assembly rules
The FG convergence greater than
would be expected from random
formation of trait groups
Fukami et al, 2005
Mechanisms
• Environmental influence: Productivity
– High Productivity, where competition is important, more rapid trait
convergence (Grime 2007)
– High productivity, greater potential for priority effects, more
divergence (Chase 2003, Fukami and Lee 2006)
• Lower species pool at low productivity = convergence
• As productivity rises, invading species can alter resource environment
or change predator density, which can create positive feedbacks
through priority effects
• Unclear if at high productivities there can be a shift back to
convergence
Chase 2003
Effect of nitrogen on convergence
varies with productivity
average pairwise jaccard
distance (dissimilarity) in
species composition among all
replicate plots within the
treatment
N fertilization often causes
divergence (positive ln RR)
Less productive communities
are more likely to diverge
Chalcraft et al, in press, Ecology
• Initial convergence, then
divergence
• From emergence to seed
set, high N does not
diverge while low N and
ambient diverge.
• Traits of the dominant
species? Amsinckia
shows positive frequency
dependence
Dissimilarity among plots (ED)
Convergence and N availability
Low N
Ambient
Rebecca Aicher, UCI
High N
Initial
Emergence
Peak season
Seed set
Mechanisms
• Environmental influence: Disturbance
– High disturbance, more regeneration which is stochastic, trait
divergence (Grime 2007)
– High disturbance, more convergence (Chase 2003)
• Less species of the regional pool can persist in disurbed ares
compared to undisturbed
• Lower densities make positive feedbacks through species effects less
likely. Species with colonist traits will be less likely to preclude
establishment by other species through interspecific interactions.
Chase 2003
Disturbance: drought decreases
stochastic community assembly
All ponds assembled in identical
environmental conditions, after 2
years ½ had drought treatment
Ponds that experienced a drought 2
years before were very similar in
species composition – deterministic
Ponds lacking drought more
dispersed, indicating larger role of
contingencies
Chase, PNAS 2007
Filters control both priority
effects and trait distribution
• disturbance filter
– Reduction in disturbance increases
importance of competition
– Increase priority effects, alternative
states
– Increased trait over-dispersion (trait
variability among species, trait-group
composition divergence)
• productivity filter
Decreased disturbance or increased
production will lead to trait
divergence
Fukami and Lee 2006, Oikos
trait convergence in the established
phase and trait divergence in the
regenerative phase
• productivity filter
– convergence in traits related to the physical and
chemical processes
– These are “effect” traits that drive ecosystems: dry
matter production, carbon storage, nutrient cycling,
anti-herbivore defense and litter decomposition.
• disturbance filter
– divergence in regenerative and phenological traits.
– these are more likely to be “response” rather than
“effect” traits.
Grime, 2006, JVS
Relationship between effect
and response traits
• If response and effect traits are not
correlated, expect change in ecosystem
function to be highly variable
• If they are correlated, then change may
be either dampened or accelerated
• If uniformly distributed, system should be
most resilient.
Suding et al. 2008, GCB
Chaos!
Elisa Benincà et al, Nature 2008.
Synthesis
• Traits may be most predictive of simple
deterministic dynamics
• Distinction between random divergence and
divergence to two or more “states”
• Complex contingencies may be very hard to
predict with traits
– Process strength (systems with strong plantsoil feedbacks)
– Different types of traits, species packing
(complementarity, relative species effects)
• Trade-off between establishment and
regenerative traits