Information about the wind speed and wind direction is of great importance for an airplane during take-off or touch-down. When pilots have this information at their disposal, they can use this to take off or touch down safely. In other cases can be decided to wait for a while before landing or taking off. When we speak about landing, in certain situations even can be decided to flight to another airport. For this reason windmasts are placed at airports nearby the several take-off and touch-down zones, so that the information required can be obtained.
This paper handles a statistical procedure that makes a prediction about the 2 minute average wind speed at a certain windmast. When this windmast temporally doesn’t function (for example due to a defect or to maintenance) through which he isn’t able to generate measurement data, the developed procedure can be used for approximating the wind speed at this windmast. For this same goal, nowadays airports often make use of back-up schemes by which values of wind speed and wind direction that are measured at a certain windmast are used in substitution for the wind speed and wind direction at the windmast which is out of order. This method has three important disadvantages. Firstly, due to differences in terrain roughness, a substitution can be severely biased. Secondly, making use of merely one other windmast, this method doesn’t take account for the capricious pattern that characterizes wind speed and wind direction and at last the back-up scheme doesn’t say anything about the reliability of a substitution.
The procedure discussed in this paper is based on a regression model that measurements, obtained from more then one other potential windmast at the airport, uses to make a prediction about the wind speed at a windmast which is out of order. Hereby the capricious pattern of the wind speed has been smoothed out. Furthermore the model deals with the several differences in terrain roughness at the airport. This has been done in two different manners. At one side by clustering to wind direction (discrete manner of modeling) and at the other side by making use of Fourier analysis (continued manner of modeling). Both methods create coefficients that are dependent of the wind direction. It has been shown that modeling with the help of Fourier analysis satisfies the most in favour of clustering to wind direction, by which is corrected fully for deviations in wind speeds between windmasts caused by differences in terrain roughness.
The procedure is also provided with an error margin that stands for the extent of the reliability of a certain prediction. This reliability appears to be strongly dependent of the state of the weather at the moment of making the prediction. The length of the error margin has been adjusted to this, so that the procedure takes into account phenomena as heavy thunderstorms but also calmer weathertypes with small changes in wind speed.
Because of modeling terrain roughness and giving error margins the procedure is useful for application during extreme weather conditions. This is an important advantage of the procedure in favour of operational back-up schemes, because of the great importance of adequate information about the wind speed during extreme weather conditions. The latter is the main reason to advice replacing the operational back-up schemes at airports by the method discussed in this paper.
Ilja Smits. Back-up modellering van windmeetmasten op luchthavens
KNMI number: TR-220, Year: 1999, Pages: 71