engleski

Evaluation of the analog-based wind speed forecasts during strong wind episodes in the coastal region of Croatia

In the coastal region of Croatia, wind speed often reaches gale values during the bora and jugo wind episodes and can cause damage to people and property. Thus, the need for correct and timely wind speed forecasts is evident. At the Croatian Meteorological and Hydrological Service, besides the raw wind speed forecasts calculated using the operational model ALADIN-HR, we use the analog- based method for point-based postprocessing. Two variations of the method are presented: the one with optimized predictor weights (AnEnT) and the other with both optimized predictor weights and correction for high wind speed (AnEnK). The optimization process quantifies the impact of each predictor variable on the similarity search process in the analog method, leading to better analogs. Additionally, since the post- processing methods can be prone to underestimate extreme events, the correction for high wind speed is added to AnEnK forecast when raw NWP forecast exceeds a certain value. For the cases tested, the large errors occur mainly due to short-term weakening of the wind-speed within the strong wind episode, or because of the time delay of the beginning or the ending of the forecasted strong wind episode. The results show that the AnEnT and AnEnK methods outperform raw NWP in general, as well as for various range of cases. Finally, during the episodes of strong bora and jugo wind, the AnEnK method often brings the predicted values closer to the observed ones, thus fulfilling the main goal of correction applied.

post-processing, analog method, strong wind, coastal region

nije evidentirano