The development of frameworks that account for community stability and its loss to environmental disturbance (e.g. regime shifts) is central to ecology, particularly for reducing uncertainty of ecological change in increasingly variable environments. Notably, community responses to disturbance often appear abrupt and surprising, raising concerns for our ability to anticipate and manage such regime shifts. In this thesis, I explore the conceptual model that compensatory dynamics may negate the effects of disturbance prior to community restructure (i.e. changes in species composition) and that their recognition may advance our ability to anticipate loss of stability. I examine the idea that the failure to recognise the weakening of mechanisms of resistance to intensifying disturbance underpins the surprise of regime shifts. My assessment centred on a plant-herbivore interaction (herbivorous gastropods-turf algae) that counters the loss of kelp forests to competitors (turf expansion) as driven by abiotic disturbances that coalescence across multiple scales of space (global to local) and time (gradual to abrupt). My tests of the hypothesis that herbivores negate the positive effects of abiotic change on turf production suggested that ecological systems might compensate for disturbance via mechanisms that prevent structural changes. Whilst global (carbon enrichment) and local abiotic change (nutrient enrichment) may drive shifts in ecological systems by altering dominance relationships between competing species (e.g. shifts from kelp- to turf-dominated reefs), adjustments in strength of herbivory appeared to negate such change. My tests suggested that resistance to change may result from the aggregate effects of individual responses (per capita consumption) where these generate dynamics that prevent change in community processes (productivity). Such dynamics may be underpinned by the necessity of individuals to maintain homeostasis in varying environments. Critically, combinations of gradual (warming) and abrupt ...