Statistical Analysis of Wind Components

Abstract: In the world of aviation, the need for accurate weather data is very important in order to plan and conduct a safe flight. Due to technology, there are ways to recieve accurate weather data live today, but also, due to technology, this data might not be available. Thus the need for accurate statistical data is desirable. Is this possible to achieve using the SADIS data we currently have available? We will try to find the answer to this question by modelling the data as either an AR(p) or an ARMA(p,q) model using different types of estimation. It turns out that we can model the wind speed rather well using an AR(3) model and the wind direction using an ARMA(5,7) model, where the parameters are estimated using recursive LMS. In conclusion we can see that the type of data we have can be modelled using straightforward methods. However, we cannot simply take our predicted models and apply those everywhere. But what we can do, is take the methods which resulted in the given models and apply that to all desired locations.