Linking theory and experiments: A meta-analysis of multi-trophic warming experiments

Background/Question/Methods

Explicitly relating theory to empirical results is a familiar challenge in consumer-resource research. However, concern for climate change impacts and an interest in metabolic ecology have renewed efforts to integrate experimental results and theory for how warming affects trophic interactions. Experiments have shown that temperature changes can alter relative abundance, stability and productivity of simple food webs, but results are highly varied, and difficult to relate to long-term projections or to other experimental systems. Recently developed theory on the relationship between temperature and trophic interactions provides one avenue for identifying when divergent experimental results are consistent with a common set of mechanisms. We reviewed published warming experiments to test two questions: 1) Are emergent trends and patterns in multi-trophic warming experiments consistent with theoretical predictions based on temperature-dependent consumer-resource interactions? 2) How can experiments most effectively contribute to a general framework for how warming affects consumer-resource interactions? We used a temperature-dependent consumer-resource modeling framework based on asymmetries in the thermal dependence of rate parameters to identify experimental conditions that could test theory. Next, we searched the literature for all warming experiments that measured rates or abundances of trophically interacting species. Finally, we tested theoretical predictions against empirical data. 

Results/Conclusions

We used theory to identify thresholds for and ratios of model parameters that facilitate predictions for how a change in temperature should affect a system. This approach generates predictions for how a system will change, these theoretical solutions require less information that would be required to fully fit a model to data. We found that most published, multi-trophic warming experiments are suited to testing qualitative, but not quantitative, predictions for how warming affects consumer-resource interactions. The most common limitations were: insufficient information reported for a) response variables to test model outcomes and b) experimental conditions to determine whether they met assumptions implicit in the hypothesis that temperature directly affects consumer-resource interactions. We suggest that warming experiments more effectively test and advance a coherent framework if they: a) consider theoretical metrics and related assumptions, b) test more than two temperatures, and c) interpret results within the context of temporal dynamics of the system. A theoretical framework to guide the design and interpretation of warming experiments would enhance the contribution of experimental findings to our understanding of relationships between metabolism and foodwebs, and increase their relevance to projecting long-term effects of warming on ecological systems.