Resilience is an emergent property of systems - and is the amount of disturbance a system can absorb before it crosses a critical threshold and reorganizes. Reorganization can return a system to its original state (undergoing an adaptive cycle, see right), or the system can reorganize into a different state.
Frequently, new states of systems are not desirable, because humanity adapted to the systems we inhabit and relies on a relatively predictable suite of ecological goods and services, which changes when a system reorganizes into something new. An example of an undesirable reorganization following collapse is the sudden onset of the Dustbowl following relatively slow landscape change and drought. The recovery of the landscape affected by the Dustbowl also provides a positive example of transformation - the purposeful erosion of the resilience of a system in an undesirable state and human-mediated transformation to a more desirable state.
Ecological resilience and engineering resilience are often used interchangeably but are very different. Engineering resilience (also called resiliency or bounceback), is the rate of return of a system following disturbance. This is sometimes a useful metric, and is included in the concept of ecological resilience, but bounceback fails to account for thresholds in systems (it assumes a system can always bounceback given sufficient time), and the fact that systems are always changing (i.e., are non-stationary).
Crawford Stanley Holling is considered the father of resilience research and crucial to the establishment of resilience research centers around the world. Holling highlighted the importance of considering surprise, system reorganization, and learning when trying to understand social-ecological dynamics. These efforts lead to new ideas about the dynamic nature of resilience and the co-development of Adaptive Environmental Management and Assessment, a learning-based approach to the management of complex environmental problems.