A framework for community interactions under climate change Sarah E. Gilman 1, 2 , Mark C. Urban 3 , Joshua Tewksbury 4 , George W. Gilchrist 5, 6 and Robert D. Holt 7 1 Joint Science Department, Claremont Colleges, 925 N. Mills Ave, Claremont, CA 91711, USA 2 Friday Harbor Laboratories, University of Washington, 620 University Road, Friday Harbor, WA 98250, USA 3 Department of Ecology and Evolutionary Biology, University of Connecticut, 75 N. Eagleville Road, Storrs, CT 06269, USA 4 University of Washington, Department of Biology, 24 Kincaid Hall, Box 351800, Seattle, WA 98195, USA 5 Department of Biology, Box 8795, College of William & Mary, Williamsburg, VA 23187-8795, USA 6 Division of Environmental Biology, National Science Foundation, 4201 Wilson Blvd., Suite 635, Arlington, VA 22230, USA 7 111 Bartram Hall, Department of Biology, University of Florida, Gainesville, FL 32605-8525, USA Predicting the impacts of climate change on species is one of the biggest challenges that ecologists face. Pre- dictions routinely focus on the direct effects of climate change on individual species, yet interactions between species can strongly influence how climate change affects organisms at every scale by altering their indi- vidual fitness, geographic ranges and the structure and dynamics of their community. Failure to incorporate these interactions limits the ability to predict responses of species to climate change. We propose a framework based on ideas from global-change biology, community ecology, and invasion biology that uses community modules to assess how species interactions shape responses to climate change. Climate change and species interactions Global climate change affects natural [1,2] and human- modified [3,4] environments, and is progressing faster than previously recorded [5]. Predicting and ameliorating the consequences of climate change presents a major challenge for ecologists, but we have largely failed to generate com- prehensive predictions or clear suggestions for ameliora- tion. We argue that this failure stems from the lack of a strong theoretical and empirical foundation for incorpor- ating species interactions into climate- change predictions. Species interactions are among the most important forces structuring ecological communities and are com- monly climate-dependent [6–8]. Davis et al. (Box 1) ele- gantly demonstrated that sets of species respond differently to temperature if reared together rather than alone. More recently, a review of >600 studies [9] found that climate change influenced virtually every type of species interaction, yet most models of climate effects on species ignore biotic interactions. It is unsurprising that many such models produce inaccurate results [10–12], but often improve if species interactions are incorporated [12,13]. We believe that a ‘community modules’ approach (sensu Ref [14], see Glossary) that simplifies entire food webs into tractable sub-webs of strong interactors can help elucidate how species interactions will influence (and be influenced by) climate change. Direct and indirect effects of climate change Throughout this contribution we emphasize the effects of temperature change, but our ideas pertain equally to other aspects of climate change such as precipitation, atmos- pheric carbon dioxide, and ocean pH. Temperature can directly cause extinctions or geographic range shifts, but Opinion Glossary Apparent competition: A three-species interaction in which two prey species negatively affect each other’s fitness through interactions with a shared predator. Boltzmann factor: The quantity e E=kT where E is the activation energy of a reaction (broadly defined to include interlinked sets of reactions, such as biological growth processes), T is temperature (in Kelvin), and K is Boltzmann’s constant. Climate envelope model: An algorithm that predicts how climate variables relate to the present or future distribution of a species. Community module: A small number of species (2–6 species, often interacting strongly) encapsulating a commonly encountered configuration of interac- tions, such as a predator–prey or host–pathogen pair, keystone predation, or a trophic cascade. Ecological release: Increased fitness of a species owing to the lack of negative species interactions, often after a range expansion or invasion that leaves some negatively interacting species behind. Facilitation: The increase in fitness of one species owing to positive interactions with another species. Keystone predation: A multi-species interaction in which a predator mediates the coexistence between two asymmetrically competing species by selectively preying on the competitive dominant, permitting persistence of the subordinate. Leading range margin: The edge of a species distribution where abiotic factors are changing to favor its expansion; usually in the direction of the poles or higher altitudes because warming temperatures allow species to colonize previously unsuitable colder habitats. Parapatric distribution: Two species with contiguous but non-overlapping ranges. Priority effect: When an early-arriving species gains an advantage over late- arriving species. R*: The resource level necessary for a species to persist in a location. If two species are in competition for a single limiting resource, all else being equal, the one with the lowest R* will exclude the other, at least in simple models of exploitative competition (e.g. with no unstable dynamics). Realized niche: The n-dimensional set of abiotic and biotic factors (including species interactions) that determine the positive rate of increase of a species. SIR model: An epidemiological model of a pathogen in a host population that tracks the density of host individuals that are susceptible (S), infected (I), or recovered (R). Trophic cascade: The effect of one higher trophic level on a lower one through an intermediary trophic level. Corresponding author: Gilman, S.E. (sgilman@jsd.claremont.edu). 0169-5347/$ – see front matter ß 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.tree.2010.03.002 Available online 12 April 2010 325