Wake vortices are the largest single component giving rise to restrictions at busy airports by way of the standard separations that are imposed. These separations are set by the International Civil Aviation Organisation (ICAO) and are set using aircraft weight categories with various separation distances imposed depending upon the relative weight categories of any particular aircraft pair. The separation distances have been calculated using both theoretical information and observed data obtained during the 1970s although there is no dependence on the prevailing meteorological conditions included in these calculations. It is commonly believed that these separation standards are very conservative and that they could be safely reduced although it has proved to be very difficult to support this belief with hard facts despite many years of research across the world. One component of any such case would be to quantify and improve techniques for predicting wind vectors over very short time periods and distances. This could lead to a system of variable separations depending upon both the aircraft weights and the meteorological conditions.
This paper describes attempts to predict one-minute mean crosswind components over short periods of at most twenty minutes. It is known that the use of persistence (assuming that the wind speed will remain constant from one minute to the next) gives a useful forecasting technique over such short periods. However, it has been demonstrated that the level of accuracy from such forecasts is only marginally useful when attempting to predict the motion and decay of aircraft wake vortices.
The prediction techniques used are purely statistical in nature although some meteorological knowledge is applied to some of the outcomes, i.e. results are not applied unless they are sensible in a meteorological context. The main method used is that of an autoregressive, integrated, moving average (ARIMA) process of which several alternatives are suggested and compared.
The results of the study show that a straightforward ARIMA process can reduce the standard errors of a prediction using persistence by around 15% with the possibility of future enhancements increasing this figure to 25%
The 8th Conference on Aviation, Range, and Aerospace Meteorology