The concept of ecological resilience embodies both an appreciation for understanding complex systems and the importance of nonlinear dynamics. This systems emphasis is a departure from reductionist traditions in biology and call for the development of methods that permit theory-based, yet operational approaches for discovering the specific attributes and mechanisms that generate resilience. Without a method for discovering the specific factors that influence resilience, our science remains of limited applicability. In this paper I illustrate the use of the graphical modeling methodology called structural equation modeling (SEM) for discovering the multivariate factors and conditions that regulate resilience in ecological systems. I begin with theory formalization and then present a meta-model as a general hypothesis about how resilience relates directly and indirectly to (1) elements of ecological memory, (2) system responses to changes and disturbance, (3) degrading agents, and (4) historical influences. The application of this meta-model is illustrated using data from a study of responses to fire by endangered grasslands being invaded by exotic trees.
Results/Conclusions
Results from this analysis show resilience to be enhanced by a high gamma diversity/species pool, high community biomass, dominant grass identity, high fire severity, and growing-season fire. Resilience in this system is impaired by increasing amounts of the degrading agent (Chinese tallow trees). It was found that resilience was strongly conditional based on location along an environmental (microelevational) gradient. This information is used to project the outcomes of various scenarios to estimate the combinations of conditions that will or will not support system resilience.
