COS 61-5
Predicting effects of community structure on ecosystem functions across trophic levels
Many ecologists have used the response/effect trait framework to help understand ecosystem functioning. This framework predicts ecosystem functions using covariation between species’ response-traits, which respond to environmental gradients, and species’ effect-traits, which have effects on ecosystem functions. However, the response-effect framework requires knowledge of species abundances to scale from individual-level traits to community-level effects on ecosystem functions. Species abundances are difficult to predict, especially across all species in a community, which can introduce considerable uncertainty to the framework’s predictions. This problem might be solved by trait-based community assembly models that have recently advanced our ability to predict community structure (ie. relative species abundances). In this study, I asked whether the Community Assembly by Trait Selection (CATS) model can predict community structures that reflect observed relationships between communities and ecosystem functions. My approach consists of a response-stage that uses CATS to predict community structure across environmental gradients, and an effect-stage that uses community structure to measure effects on ecosystem functions. I used data I collected on abundances of herbivore and predator functional groups in a Colorado meadow. My objectives were to predict predator community structure over a prey resource gradient, and the effects of predator traits on ecosystem functions.
Results/Conclusions
The results indicate that merging the response/effect framework and community assembly modeling may be useful for anticipating ecosystem functioning. The CATS model predicted much of the variation in predator relative abundances across a multivariate herbivore gradient (Hellinger h = 0.77). Also, the predicted community structures yielded community-level trait measurements that had similar relationships with ecosystem functions as the observed traits. Both observed and predicted community-weighted mean trait values and trait diversity (ie. functional diversity) were correlated with tertiary production (an ecosystem function). Additionally, I am using this dataset to apply my approach at multiple scales of study; further analyses are needed to evaluate the general usefulness of this approach. Relationships among community-level trait metrics, and between traits and ecosystem functions, are often complicated. These relationships may be elucidated by considering how community structure is governed by community assembly. Because knowledge of community structure is essential for calculating the community-level trait metrics used in the response/effect framework, new community assembly methods may aid efforts to predict ecosystem functioning.