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 acceptably high probability of detection and low false alarm
rate, especially in clear air at upper levels.
This paper discusses ongoing efforts at the National Center for Atmospheric Research (NCAR) to forecast turbulence in clear air at upper levels. The technique used integrates observations of turbulence from pilot reports and ACARS equipped aircraft (vertical accelerometer and eddy dissipation rate data) together with turbulence diagnostic algorithms based on RUC2 numerical weather prediction model output. A single turbulence diagnostic algorithm is not used, but rather a fuzzy logic combination of many different commonly used turbulence diagnostic algorithms.
Forecast statistics compiled for several case studies over the continental U.S. are presented which demonstrate this approach is capable of substantially improving upper level turbulence forecasts over any single turbulence diagnostic when used in isolation, in fact to the point where the probability of turbulence detection and false alarm rate are operationally useful.
The 8th Conference on Aviation, Range, and Aerospace Meteorology