The Outlook: Scattered Showers

This is a philosophical article that discusses turbulence, chaos, and limited predictability as they affect professional meteorologists. Particular consideration is given to the problem of the limits of predictability in meteorology. It is a science whose behavior is not regular; it is subject to the theory of chaotic dynamics of nonlinear systems as pointed out by Lorenz. Automation and computerization can be perfected and used to forecast weather routinely for 3-5 days with minimum human intervention. However, the forecaster will not be eliminated-not because human beings are necessary to make forecasts, but because there is a limit to that which can be predicted. The theory developed by Lorenz, Leith, and others gives climatological values for the useful prediction range based on an assumed shape of the energy spectrum of atmospheric motion. On the average, prediction cannot be extended beyond 14 days, even with perfect observation technology and arbitrarily large computers. For reliable guidance on predictability, a simplified subjective version of the lagged-averages forecasting proposed by Koffmann and Kalnoy is suggested. The concept of ambiguity expressed by the phrase Fourier-transform ambiguity is proposed to substitute for the negative connotation that goes with uncertainty, insecurity, indeterminability, and unpredictability.