Scenarios can help planners and decision makers to think through uncertainties about the future and make decisions that are robust to a variety of possible outcomes. To develop useful scenarios we need to understand the main processes of relevance to the system of interest.
Ecological processes, and the feedbacks that they can create between human actions and human well-being, are thought to be important for human societies. Current uncertainties over the long-term resilience of ecosystems and the substitutability of ecosystem goods and services can be translated into three alternative realities: ecosystems may be relatively brittle, relatively resilient, or largely irrelevant. Although these extremes are only rough characterizations of reality, they help us to focus our thinking about the possible outcomes of interactions between humans and the rest of the biosphere. Existing global scenarios can be categorized into a small number of families based on shared themes and assumptions about the future. Considering the internal consistency of four of the main scenario families in relation to the three alternative ecological realities suggests that all existing scenarios make strong, implicit assumptions about the resilience of ecosystems. After a detailed discussion of individual examples, we present a synthesis of the incorporation of ecology in existing scenarios. All current scenarios are inconsistent with at least one possible property of ecosystems and their likely interaction with society. The interrelationships between ecological reality, human views of ecosystems, and social responses to actual and perceived ecological change are complex. For the Millennium Ecosystem Assessment and future scenario exercises, we recommend that essential ecological assumptions should be made explicit to ensure that the details of each scenario are consistent with both the perceived and the actual degree of resilience of ecosystems.