Predator-prey interactions in natural communities are complex, with predators often exploiting multiple prey types and generating indirect interactions among them. Ecological theory has traditionally modeled these interactions using functional responses models which are based on foraging rates, not energy transfers. This approach overlooks how the energy acquisition rate of a predator can alter its behavior and, in turn, the strength of species interactions.

Here, we integrate predator energetics into a functional response model to represent trade-offs predators face when foraging on prey that vary in risk and abundance across heterogeneous landscapes. We compared model predictions to 20 years of prey species density and reproductive success data. The mechanistic model was parameterized for an Arctic tundra vertebrate community, where the Arctic fox feeds on cyclic lemmings and eggs of sandpipers (non-risky prey) and gulls (risky prey that often nest in partial refuge like islands). In this system, predator-mediated interactions generate apparent mutualism between lemmings and birds, but its strength varies between species, and the mechanisms underlying this interaction remain unclear.

We found that fox energetic balance was highly related to lemming density, with a threshold of 89 lemmings per km² required for a positive energetic balance. Model-predicted gull nest acquisition rates were lowest on islands when the energetic balance of foxes was positive, and highest for nests on the shore when foxes were in deficit. The model that incorporated predator risk-taking behavior and energetic balance produced variation in gull hatching success that most closely matched empirical observations.

We documented for the first time that a shift in predator energetic balance, triggering changes in attack and capture probabilities on a risky prey, can be a key mechanism underlying the apparent mutualism between lemmings and gulls. In contrast, for non-risky prey, the indirect effect can be essentially driven by changes in predator movement. These findings highlight how prey characteristics can lead to different mechanisms behind similar indirect interactions.

Taken together, our results indicate that mechanistic models integrating species traits, landscape features, and energy-dependent behavioral adjustments can improve our ability to quantify interaction strengths in natural communities.