Space-Time Modelling with Long-Memory Dependence: Assessing Ireland's Wind Power Resource

Abstract

We consider estimation of the long term average power output from a wind turbine generator at a site for which few data on wind speeds are available. Long term records of wind speeds at the 12 synoptic meteorological stations are also used. Inference is based on a simple and parsimonious approximating model which accounts for the main features of wind speeds in Ireland, namely seasonal effects, spatial correlation, short-memory temporal autocorrelation and long-memory temporal dependence. It synthesizes deseasonalization, kriging, ARMA modelling and fractional differencing in a natural way