Evolutionary Assembly of Food Webs


My research on predator-prey interactions and habitat selection is generally focused on consumer-resource relationships at short time scales and small spatial scales.  However, I am also interested in questions about the robustness of these relationships to perturbations (e.g., changes in habitat characteristics, species’ abundances, or primary production) and in the larger scale consequences of adaptive behaviors (and responses) of predators and prey.  To investigate these questions, I have been developing a simulation model of food web assembly.  In this model, each species is represented in an abstract fashion as an L-bit genotype (i.e., a string of L ones and zeros, where L typically varies between 64 and 256).  The model has three time scales: a short time scale on which predator-prey relationships are adaptively adjusted, a medium time scale on which population dynamics occur, and a long time scale on which speciation events take place.  The model thus borrows parts of several previous approaches to this topic (e.g., the “Webworld” as described by Drossel et al. 2001, and the “matching” model described by Rossberg et al. 2006).  Algorithms governing processes at the different time scales determine how the composition and structure of the food web evolve.  The advantage of using this type of model is that the properties of species and their relationships with each other emerge from a combination of natural selection and stochastic events.  Unlike many previous models, trophic relationships and food web connectance are not specified by the model.  Rather, these properties are outcomes of the implementation of simple, biologically realistic rules.

          Already this model has produced several interesting results.  For example, in the absence of perturbations or speciation events, the food webs that evolve are stable over long periods of time.  However, even with no other perturbations, the occurrence of speciation events continually triggers turnover in species compositions and occasionally triggers mass extinctions.  In other words, the system never reaches a state at which it is immune to invasion by a new species.  Currently and in future work I am using and will use this model to address several questions.  First, how robust or fragile are food webs to species deletions?  Second, what are the properties of food webs that confer robustness?  Third, how do different constraints on species’ evolution impact the properties of food webs?

samuel m. flaxman

Assistant Professor, Department of Ecology and Evolutionary Biology

University of Colorado Boulder