Climatic impact assessment is generally conducted by reference to numerical models, from which most estimates of climatic change are derived, and to the policy developers, by whom the impact assessments are demanded. The propagation of estimates derived from numerical climate model predictions of greenhouse-induced climate change through impact models into policy advice is a precariously uncertain process which compounds the considerable uncertainties already inherent in policy development. Clear statements of scientific confidence in the greenhouse phenomenon in the mid-1980s prompted demands for policy, and hence for policy advice. In Australia, as in many other countries, public and political awareness of the possibility of greenhouse-induced climatic change increased. These developments led to the formation of the Intergovernmental Panel on Climate Change (IPCC); to the Framework Convention on Climate Change, signed at the United Nations Conference on Environment and Development (UNCED) in June 1992; and to the review of the World Climate Programme in April 1993. This special issue of Climatic Change illustrates some aspects of the difficulties surrounding projections of climatic impacts at a national scale where policy development almost always occurs under conditions of uncertainty. It may be valuable to identify uncertainty issues which could benefit from additional research and also sensitive points in the policy development process at which uncertainty can be used and abused. In this paper, the role of uncertainty in the greenhouse debate is reviewed from the perspective of a natural scientist working in a developed country. The aim is to offer a framework for the rest of this special issue of Climatic Change. Uncertainty is by no means the only factor which influences views on climate change but increased understanding and more informed debate of all aspects of the uncertainties relating greenhouse-induced climatic change to policy development and implementation would be beneficial.