Environmental perturbations can strongly affect community processes and ecosystem functions by acting primarily as a subsidy that increases productivity, a stress that decreases productivity, or both, with the predominant effect potentially shifting from subsidy to stress as the overall intensity of the perturbation increases. While perturbations are often considered along a single axis of intensity, they consist of multiple components (e.g., magnitude, frequency, and duration) that may not have equivalent stress and/or subsidy effects. Thus, different combinations of perturbation components may elicit community and ecosystem responses that differ in strength and/or direction (i.e., stress or subsidy) even if they reflect a similar overall perturbation intensity. To assess the independent and interactive effects of perturbation components, we experimentally manipulated the magnitude, frequency, and duration of wrack deposition, a common stress-subsidy in a variety of coastal systems. The effects of wrack perturbation on salt marsh community and ecosystem properties were assessed both in the short-term (at the end of a 12-week experimental manipulation) and long-term (6 months after the end of the experiment). In the short-term, plants and associated benthic invertebrates exhibited primarily stress-based responses to wrack perturbation. The extent of these stress effects on density of the dominant plant Spartina alterniflora, total plant percent cover, invertebrate abundance, and sediment oxygen availability were largely determined by perturbation duration. Yet, higher nitrogen content of Spartina, which indicates a subsidy effect of wrack, was influenced primarily by perturbation magnitude in the short-term. In the longer term, perturbation magnitude determined the extent of both stress and subsidy effects of wrack perturbation, with lower subordinate plant percent cover and snail density, and higher Spartina nitrogen content in high wrack biomass treatments. However, stress effects on the marsh community were generally less pronounced 6 months after the wrack perturbation, indicating capacity for recovery. Our results demonstrate that individual perturbation components can determine the degree to which its effects on the community elicit primarily stress- and/or subsidy-based responses. Further, the nature and extent of stress-subsidy effects can change over time, depending on species' relative ability to tolerate and/or recover from perturbation.