In response to growing global impact on ecosystems, we design programs for conservation and restoration to maintain and enhance biodiversity, productivity and ecosystem resilience. To ensure the greatest return for these programs, there is an implicit requirement for identifying and understanding the complex non-linear relationships that can exist between impact gradients, ecosystem structure and the processes that mediate the two. Ecological resilience theory has developed as one of the fundamental explanations of this complexity. The application of ecological resilience theory in a local management context, however, is often hampered by a disparity between the theory and what is practical to test empirically. This thesis used seagrass ecosystems in Moreton Bay, Queensland as a model system for the development and testing of a practical framework to examine the potential for incorporating measures of feedback loops into the empirical assessment of resilience. I focussed on the behaviour of feedback processes in relation to changing levels of impact with a view to developing a generic, testable hypothesis that could be used to assess ecological resilience more broadly. A Bayesian network was used to synthesis the known relationships between impact and seagrass response to identify three key feedback processes that stabilise seagrass ecosystems in Moreton Bay......