Data and code for: Warming effects on population sizes and top-down control are dependent on temperature range.

Rising global temperatures are expected to affect natural ecosystems, but predicting the consequences is challenging given the inherent complexity of ecological communities. Natural communities are complex networks of interacting species, and temperature changes can thus influence species directly (e.g. alter metabolic and demographic rates) and indirectly via changes in their various interactions with other community members within (competition) and across (predation) trophic levels. While previous studies indicate that both predation and competition can be temperature specific, we still have a poor understanding of how both interact along a thermal gradient. In addition, temperature effects appear contradictory across studies, potentially because they are typically limited to few temperatures which cannot capture the expected non-linearity of the interactions across a thermal gradient. Here we examine the effects of temperature on a food-web including one predator feeding on two prey species that share a common resource to tease apart its direct effects on populations, indirect effects on competition and predation, and their interactions. Specifically, we experimentally manipulated the presence of a predator and potential competitor for two different prey species. We then exposed the resulting six species treatments to a temperature gradient from 10-31°C for ~3 prey generations and quantified prey population growth rate, probability of extinction and coexistence. We found that the probability of extinction, final population growth rate, and the potential for prey coexistence all demonstrated significant non-linear, hump-shaped responses to temperature. Predators significantly increased the probability of extinction for both prey, and for one prey species this effect was strongest at intermediate temperatures. Predators also significantly reduced the population size of both species. For one prey, this effect strengthened at warmer temperatures, causing prey populations to peak at lower temperatures in the presence of predators. Contrary to expectations, presence of a second prey had either no or a positive effect on prey persistence and population size, especially in the presence of predators, suggesting apparent mutualism. Whether this effect was contingent on temperature was specific to prey species and response variable. Overall, these results reveal that temperature effects on ecological dynamics are often non-linear (unimodal) and contingent on the community structure, which helps explain the context-dependent effects of warming.