Theory of temperature‐dependent consumer–resource interactions

Changes in temperature affect consumer–resource interactions, which underpin the functioning of ecosystems. However, existing studies report contrasting predictions regarding the impacts of warming on biological rates and community dynamics. To improve prediction accuracy and comparability, we develop an approach that combines sensitivity analysis and aggregate parameters. The former determines which biological parameters impact the community most strongly. The use of aggregate parameters (i.e., maximal energetic efficiency, ρ, and interaction strength, κ), that combine multiple biological parameters, increases explanatory power and reduces the complexity of theoretical analyses. We illustrate the approach using empirically derived thermal dependence curves of biological rates and applying it to consumer–resource biomass ratio and community stability. Based on our analyses, we generate four predictions: (1) resource growth rate regulates biomass distributions at mild temperatures, (2) interaction strength alone determines the thermal boundaries of the community, (3) warming destabilises dynamics at low and mild temperatures only and (4) interactions strength must decrease faster than maximal energetic efficiency for warming to stabilise dynamics. We argue for the potential benefits of directly working with the aggregate parameters to increase the accuracy of predictions on warming impacts on food webs and promote cross‐system comparisons. Consumer–resource interactions underpin the functioning of ecosystems and can be strongly dependent on environmental conditions. To improve predictions on the impacts of environmental change on these interactions, we develop a dual approach to (1) identify the most important processes driving the response (sensitivity analysis) and (2) increase explanatory power of theoretical analyses (aggregate parameters). We implement this approach on temperature‐dependent consumer‐resource interactions, and demonstrate how it can help reconcile existing mixed predictions and increase the accuracy of future predictions.