Accuracy of mean wind climate predicted from historical data through wind LIDAR measurements.

Prediction of mean and extreme wind climate at a site from historical data contains intrinsic errors; these come from the not necessarily good quality of the original data, from the approximations adopted when defining orography and roughness at the site of interest, from the possible presence of sheltering effects at the site of measurement and from the approximations embedded in the mathematical models used for transferring the measured data to the site of interest. In this paper, first the mean and extreme wind climate evaluated at three neighbouring sites is compared, and an attempt is made to relate the differences to the characteristics of the sites of measurement and to those of the dataset. Then the wind velocities and wind resource projected to a fourth site are compared among each other and with wind LIDAR measurements taken at the site of projection; this was done with the purpose of quantifying inaccuracies and possibly detecting their causes. It is found that the mean wind climate and wind resource predicted at a site of interest is indeed dependent on the original dataset and on the projection technique. In particular, a simple projection technique is proposed, based on interpolation of the data simultaneously measured at the site and at a neighbouring station. The performance of the proposed technique is compared with that obtained using a commercial software, and no evidence was found of one being more accurate than the other. This suggests that the projection technique based on interpolation could be calibrated on short-term on site measurements, and then used to assess the mean and extreme wind climate at the site of interest using long term records from the neighbouring stations.