Impacts of habitat fragmentation
on plant and insect communities:
beyond species richness!
Seminar outline
• Part 1
Using traits to understand impact of habitat fragmentation on plant
communities: local vs. dispersal processes
• Part 2
Impact of habitat fragmentation on changes in relative abundance
of flower-visiting insects
Using traits...
1. Theoretical predictions
2. Selection of traits linked to clear ecological hypotheses:
“Traditional approach”
Response
Environmental
change
Change in trait Composition (e.g.
weighted mean)
“Our approach”
Response
Trait group A
(e.g. Mobile species)
Environmental
change
Change in species diversity
Test interactions between traits
Trait group B
(e.g. Sedentary species)
Change in species diversity
3. Test using large scale datasets
How does the trait modify the response to the environmental change?
Impact of habitat fragmentation
on plant communities:
local vs. dispersal processes
Marini L., Bruun H.H., Heikkinen R.K., Helm A., Honnay O., Krauss J., Kühn I.,
Lindborg R., Pärtel M., Bommarco R. (in press) Traits related to species persistence
and dispersal explain changes in plant communities subjected to habitat
loss. Diversity and Distributions
Impact of grassland fragmentation on plants
Plant species
richness
Large number of studies testing area and connectivity effect on
overall plant species richness
Area
Connectivity
Metapopulation ecology has mainly considered mobile animals and therefore stressed
the importance of dispersal processes
However...
Local vs. dispersal processes
For plants, it is expected that species’ ability to both persist locally and disperse are
critical in shaping communities
Local within patch processes
Dispersal processes
Source
population
Unoccupied patch
Occupied patch
One approach to clarify this is to explore species richness responses to fragmentation
for groups of species with shared life-history traits
Starting hypotheses
Processes favouring species robustness to habitat fragmentation
Local processes
Traits
Asymmetric competition for light
Plant height (short vs. tall)
Asexual reproduction
Clonal vs. annual
Persistence in the seed bank
Persistent vs. transient
Dispersal processes
Animal (directional) vs. abiotic agen
(random)
Increase dispersal success
Seed number
(low vs. high)
Careful to avoid collinearity between traits!
AIMS
1. To test for interactions between traits: do any combination of traits
provide higher robustness to habitat fragmentation?
2. To use traits to understand the relative importance of local vs.
dispersal processes
Data
Extinction debt mostly paid in all regions [Krauss et al. (2010) Ecol. Lett.]
Homogenization of taxonomy and plant life-history traits across regions
Orthogonal gradients in area and connectivity (Hanski connectivity index in
all regions)
Methods: Mixed model approach in two steps
I. Testing ecologically meaningful interactions between traits
Species richness~ Trait A*Trait B*Area, random=~1|country/site
II. Testing interactions between single traits and area (or connectivity)
Species richness~ Trait*Area, random=~1|country/site
Tall
Area
Annual
?
Species richness
Species richness
Short
Clonal
Connectivity
...
Results
Species richness
Negative effect of habitat loss but no effect of connectivity
Area
Connectivity
No interactions between traits
The effect of area was modified by three traits:
1. Plant height (short vs. tall species)
2. Clonality (annual vs. clonal)
3. Dispersal agent (abiotically- vs. animal-dispersed species)
4. Seed bank
5. Seed number
Results: trait effect
Plant sensitivity to habitat fragmentation
Higher sensitivity to habitat loss for:
1. Small species (low competitive ability for light)
2. Perennial clonal (trade-off between clonality and dispersal?)
3. Abiotically-dispersed species (random vs. animal directional dispersal)
Plant sensitivity to habitat fragmentation
Results match well with other recent
studies
Lindborg et al. (2012) Ecography
Plant sensitivity to habitat fragmentation
Results match well with other recent
studies
Negative Ωj implies a negative
response to habitat loss
Montoya et al. (2008) Science
Conclusions
Our trait-based analyses gain insights into the potential mechanisms leading to plant
extinction due to habitat fragmentation
The importance of within-patch local processes have been probably underestimated
in fragmentation research so far
The interaction between local persistence and
dispersal shaped plant communities
What about changes in relative abundance?
Background
Abundance-based measures:
-Evenness
-Dominance
-Species composition
-Functional diversity...
Evenness
Evenness refers to the relative contribution of each species to the total biomass
or number of individuals
Species diversity
Species richness
Species evenness
Impact of fragmentation on evenness of flower-visiting
insect communities
Marini L. , Öckinger E., Bergman K.-O. , Krauss J., Kuussaari M., Jauker B., Pöyry J., Smith
H.G, Steffan-Dewenter I., Bommarco R. (in prep.) Contrasting effect of habitat area and
connectivity on evenness of flower-visiting insect communities
Aims
Evenness
Species evenness has been used more often as a driver of
ecosystem functioning rather than as a community response
?
Fragmentation
Which are the effects of habitat fragmentation on
abundance patterns of flower-visiting insects?
Problems with evenness definition
Looseness of the mathematical definition of evenness: several indexes
with different sensitivity to changes in rare or dominant species
The choice of the metric is central in the interpretation of the ecological
relationships between environmental drivers and evenness
The most important property is the independence from species richness
Evenness profile
0.0
Diversity profile: Community A is more diverse than a community B if the diversity
profile for community A is everywhere above the diversity profile for community
B.
Baz1
Baz2
Baz3
Baz4
Baz5
-1.0
Baz1
Baz2
Baz3
Baz4
Baz5
Baz2
Baz5
Baz3
Baz1
-2.0
-1.5
E-alpha
-0.5
From the diversity Rényi profile we
derived an evenness profile
Baz4
0
0.25
0.5
1
2
4
8
Inf
alpha
Increasing importance of changes in dominant species
Background: General predictions
Evenness
Local processes promoting evenness:
-Larger habitat diversity in large patches
-Lower inter-specific competition in large
patches
Area
Evenness
Dispersal processes promoting evenness:
-Larger exchange of individuals between
patches
Connectivity
Aim: to test these predictions using a large empirical data set
Data
Ten grassland networks
(7 for butterflies and 3 for wild bees)
Proportional sampling
Transect counts
Habitat
connectivity
Patch
Orthogonal gradients in area and connectivity
Habitat area
Slope ±CI 95%
Species evenness
Results
Area
Increasing area
Increasing importance of changes in dominant species
Results
Species
evenness
Slope ±CI 95%
Connectivity
Increasing connectivity
Weaker effect for bees than for butterflies
Increasing importance of changes in dominant species
Which are the underlying mechanisms?
Area and specialization
Fragmentation modifies the specialization distribution
P<0.01
Area
% Generalist spp.
Bees (Central foragers)
% Generalist spp.
Butterflies
P<0.01
Area
Area and mobility
% Mobile spp.
P<0.01
% Mobile spp.
Same patterns for species mobility (body size)
Area
P<0.01
Area
Small patches host less sedentary species than large patches
What about connectivity?
Connectivity-evenness relationship
P<0.01
Connectivity
% Mobile spp.
% Generalist spp.
No patterns for bees
Connectivity
% Mobile spp.
% Generalist spp.
Negative relationship for butterflies
P<0.01
Connectivity
Interpaly of local and dispersal processes
Mobile and generalists
Sedentary and specialists
Local processes:
Inter-specific competition
(nesting sites, plant resources etc.)
Different local population growth
Dispersal processes:
Inter-patch movements
Interpaly of local and dispersal processes
Increasing importance of dispersal processes
Increasing importance of local processes
Small patches are dominated by generalist immigrants,
no viable local populations: minimum area threshold?
Increasing connectivity may reduce species dominance by
favoring inter-patch dispersal of sedentary and specialist species
Interpaly of local and dispersal processes
Combinations of species exhibiting true metapopulation dynamics
with species with frequent inter-patch movements
Only large patches sustain populations that
can be locally dominant
Highly complex processes underpinning abundance patterns
Conclusions
Pollinators are expected to show drastic changes in evenness
(dominance) due to several environmental pressures other than
fragmentation
Pollinator evenness is expected to be strongly related to pollination
service
We need to evaluate multiple drivers and their interactions on pollinator
evenness!