An integrated approach to clear-air turbulence prediction

Commercial and general aviation aircraft continue to encounter unexpected turbulence that is hazardous to aircraft and passengers. Accurate one to six hour forecasts of moderate or greater turbulence in a three-dimensional airspace are needed by the aviation community, but unfortunately, the small scale nature of the turbulence and the large scale nature of the existing meteorological observational network has not to date allowed turbulence forecasts with an acceptable high probability of detection and low probability of false alarm, especially in clear air at upper levels, This paper discusses efforts ongoing at the National Center for Atmospheric Research (NCAR) to integrate new observation techniques with various turbulence forecasting algorithms. The observation techniques allow timely nowcasts and provides quantitative data for forecast verifications. The forecasting technique uses fuzzy logic to merge many turbulence forecast indices together, taking into account known regions of validity (in terms of time and space, and also in terms of the details of the synoptic environment) of the various component algorithms, and weights each component accordingly to determine the final forecast.