Computer Simulated Climate Forecasts
Climate science research in the 1980’s and 1990’s devoted considerable attention to developing computer models capable of forecasting general climate trends on the basis of the information then known. Computer generated scenarios have been used to suggest specific global and regional effects of positive climate change, such as increased or decreased local precipitation, longer or drier growing seasons, and coastal inundation. At the time of the ICPP’s Second Assessment, the authors of that document were cautious regarding the accuracy of global climate forecasts, especially at the regional level. Such caution was based, in part, on the difficulties of modeling the complex atmospheric feedbacks associated with water vapor, clouds, ocean circulation, and the albedo effect. At the time of the Second Assessment, most simulations were unable to replicate short-term climatic variations, such as El Niņo, without being manipulated. Since then, computing power has improved, as have the models themselves and the instrumental data that is fed into them. When tested against current and past climate observations, current models earn a higher degree of confidence than did their forerunners less than a decade previously. The IPCC now considers climate simulations capable of providing “credible simulations of both present annual mean climate and the climatological cycle,” as well as “stable, multi-century simulations.” If a simulation, incorporating all known atmospheric feedbacks, can faithfully reproduce several centuries of recorded climate variation, the odds of the same simulation running an accurate forecast well into the future increase.