Optimization of value of aerodrome forecasts

Prediction of short-term variations of vital boundary layer conditions at airports, such as visibility and cloud base, is important to the safe and economic operation of airlines. Results of an experiment involving groups of forecasters at three different locations across Australia are presented. The forecasters were asked to indicate their degree of confidence that weather at the airport would be below "minimums'' that would require aircraft to carry adequate fuel to proceed to an alternate destination should they not be able to land. The results of the trial are shown to closely obey a Gaussian model as used in signal detection theory (SDT). The data are fitted to an accuracy-value model developed by Mason. The paper demonstrates the ability of forecasters to provide reasonably reliable probability forecasts of significant events at airports. The potential value in reliable estimation of the probability of low visibility and cloud base at aerodromes is estimated by using cost parameters for two actual examples of flights into Melbourne and Townsville, Australia.