Background/Question/Methods   The paleoenvironmental perspective informed by proxy records makes it clear that climate system change is not always as gradual as the climate forcing behind the change, but rather is often characterized by “tipping-points” and threshold responses. Thus, even though future global warming is likely to be up to an order of magnitude faster than any global warming of recent Earth history (e.g. over glacial terminations), there is a strong likelihood that many future regional changes will be even faster as the system is pushed increasingly over thresholds. For example, parts of the Antarctic Ice Sheet that are currently grounded below sea level could transform abruptly into a period of rapid (i.e., more than one meter per century) global sea level rise totaling many meters of sea level rise. Ocean currents could also shift abruptly, as could regional hydrologic regimes – a particularly likely scenario in regions that are already water limited. Evidence of such behavior is abundant in the paleoclimatic record, and must be taken into account in assessing the possible scale of future ecosystem impacts or in developing robust adaption strategies. Moreover, even gradual climate change can push ecological systems across tipping points where non-linear widespread changes in disturbance and mortality trigger large scale abrupt transformations in ecosystems and the services they provide. Such abrupt changes are already taking place in response to climate change, for example as anthropogenic warming combines with drought to create moisture stress in excess of a species' capacity to adapt. Such abrupt ecosystem change can, in turn, drive abrupt climate system feedbacks that result in even greater ecosystem impact.

Results/Conclusions   Given that abrupt change could be one of the most difficult aspects of climate change to deal with, research needs to focus increasingly on 1) understanding the underlying non-linear dynamics well enough to simulate them with climate and ecosystem models, 2) using models combined with records of past abrupt change to assess the probability of future abrupt change until models alone have the needed non-linear dynamics, 3) using both models and records of past abrupt change to explore the potential for early warning, and 4) developing the linked human-environment knowledge needed to reduce vulnerability to abrupt change that cannot be predicted with skill. The sustainability of many natural (including biodiversity) and human systems may ultimately be most threatened by abrupt change that occurs without warning and in the absence of robust human-aided coping strategies.