Use of Computers to Simulate and Predict Climate Change in the 21st Century
Authors: Charles F. Keller, Cyber Mesa, New Mexico
Abstract: There are two approaches to the usefulness of models. The more philosophical one is used by many pundits with phrases like: ―models are only numerical experiments and can never tell you exact truth.‖ This approach is fine as far as it goes, but in the end shows little appreciation for what models really do. It makes pronouncements about models‘ ability to make accurate predictions of climate change, but ignores the day-to-day usefulness of them. For, while non-modelers argue in a fairly ethereal realm, there are hundreds of scientists using models to analyze and make sense of the masses of observational data coming in from satellites, ocean buoys, balloons, surface stations, etc. Without models, there is little chance we can understand what the observations are telling us. Another thing not discussed much is the ability of models to perform ―what if?‖ experiments. What if we cut down the Amazon rainforest? What if the polar ice caps melt? What if we changed the amount or kind of aerosols we put into the atmosphere? What if the Gulf Stream shut down? What if we engineer climate? What if there were no low level clouds, a different heat flux between oceans and atmosphere, etc.? Observations can only tell us about one climate experiment. Models can examine a large number and by so doing, quantify the various forcings—determine which are important and which don‘t have much effect. There are about 20 large teams constantly improving climate models and studying all aspects of how the Earth parses its heat budget. This chapter will largely leave the ultimate questions to those looking at modeling from a safe distance and will try to look at examples of the details where real understanding and progress is occurring.