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!