This chapter assesses the ability of climatologists to provide a detailed projection of anthropogenically induced greenhouse warming suitable for use in ecological impact studies. Alternative approaches to the development of regional scenarios are reviewed. Despite their many uncertainties, general circulation models (GCMs) are considered to offer the greatest potential. The use of the present generation of GCMs, equilibrium models, in scenario development is discussed. Major uncertainties exist, particularly, with regard to the pattern of climate change at the regional level. A major reason for these uncertainties is the influence of feedback effects, many of which involve interactions between climate and ecological processes. Model results from transient GCMs are just becoming available and these suggest that the regional pattern of change differs markedly if time-dependent effects such as the thermal inertia of the ocean are taken in account. This chapter reaches the conclusion that none of the currently available models provide scenarios of sufficient reliability for use in quantitative regional impact studies. They may, however, be used to assess the sensitivity of ecosystems to climate change and in the development of impact assessment methodologies, with full awareness of their limitations.