Evaluating ecological resilience with global sensitivity and uncertainty analysis

Concern about catastrophic tipping points has motivated inquiry to better understand ecosystem dynamics in the presence of human action. This requires that we confront multiple challenges in the evaluation of complex systems. One challenge is that resilience has proven difficult to quantify; another issue is that the value of model complexity relative to system complexity is disputed; and finally, local methods for assessing uncertainty are inadequate for more complex models. We address these three challenges simultaneously by proposing a means of evaluating ecological resilience via employment of global sensitivity and uncertainty analysis and comparing models of varying complexity. We suggest that probability distribution functions in output from global sensitivity and uncertainty analysis can be interpreted in terms of ball-and-cup diagrams used in systems theory to visualize ecological resilience. This permits quantification of ecological resilience in terms of the probability of whether a system will remain in a pre-existing state or shift to a different state. We outline the methods for using global sensitivity and uncertainty analysis to evaluate ecological resilience and provide examples from recent research. We highlight applications of these methods to assessment of ecosystem management options in terms of their ramifications for ecological resilience.