A first order geometric auto regressive process for boundary layer wind speed simulation

Abstract

Abstract. Under certain conditions the first order geometric auto regressive (AR) process has statistical properties similar to atmospheric boundary layer wind speed. In this contribution, we investigate this similarity and analyse the extent to which this stochastic process is a suitable model for wind speed simulation. In particular, we focus on the fluctuation of the process around its moving average over a given number of time steps. We show that the fluctuation conditioned on the value V of the moving average are of a symmetric normal distribution with a proportionality between its standard deviation and V . This proportionality is empirically observed in wind speed data. Furthermore, we show that the increment distribution of the geometric AR(1) process is in good agreement with the symmetric Castaing distribution which is empirically found in wind speed data.