Small-scale surface wind prediction using dynamic adaptation

An efficient method for surface wind prediction based on dynamic adaptation of a low-resolution model forecast towards a finer mesh is described. The method consists of: (1) interpolation of the data field predicted by the large-scale model onto the target grid; (2) dynamic adaptation of the fields onto the new grid by running the fine-mesh model for an appropriate period of time. The time necessary for the adaptation is roughly estimated by scale analysis and by using a linear model, and confirmed by some numerical simulations. This time was found to be between 30 and 45 min for adjusting general features of the wind field; the additional time needed to adjust small-scale wind phenomena can be several times longer. It was shown that this time period does not depend on the model configuration. The main advantage of the presented method is its low computational cost. The adaptation with a simplified model is capable of reproducing small-scale features in the wind field which develop due to the new orography in the fine grid. The method fails in cases of certain atmospheric conditions or a wrong large-scale forecast.