Abstract
Nontrophic interactions can contribute to negative and positive feedbacks within a community, thus affecting likelihood of regime shifts; however, assessing the nature and importance of these effects in a network remains challenging, especially for pelagic ecosystems. Here, we present a qualitative modeling approach for assessing the importance of different effects and resultant feedbacks for community stability, using a Southern Ocean example. A potentially important positive feedback in the Southern Ocean ecosystem involves production of a chemical cue, dimethyl sulfide (DMS), by some phytoplankton. Production of DMS can promote phytoplankton growth by attracting predators of phytoplankton‐grazers, and nutrients released as feces from those predators help fertilize the water column. We explored how uncertainties in the nature of this feedback affect community stability in a set of small, community models. We found that stability varied substantially depending on how the community was modeled, but that the interactions most important for determining stability were consistent across all models. Model stability was sensitive to the strength of phytoplankton competition, controls on phytoplankton, DMS production and release, and predator attraction to DMS, suggesting that the community could be destabilized by perturbation affecting these interactions. Incorporating DMS‐mediated feedbacks into a larger Southern Ocean network had a moderate impact on stability characteristics and altered the trophic level at which the system would be most vulnerable to perturbation.