CMStatistics 2022: Start Registration
View Submission - CMStatistics
Title: Calibration of wind speed ensemble forecasts using truncated GEV based EMOS approach Authors:  Marianna Szabo - University of Debrecen (Hungary) [presenting]
Sandor Baran - University of Debrecen (Hungary)
Patricia Szokol - University of Debrecen Faculty of Informatics (Hungary)
Abstract: Probabilistic ensemble weather forecasting is an operatively used method of prediction at all major weather prediction centres. These forecasts are obtained from multiple runs of numerical weather prediction models with different initial conditions or model parametrizations. However, to account for the under-dispersive or biased nature of the ensemble forecasts, some kind of post-processing is applied. One of the most popular parametric statistical post-processing techniques is the ensemble model output statistics (EMOS), which provides a full predictive distribution of the weather quantity. We propose a novel EMOS model for calibrating wind speed ensemble forecasts, where the predictive distribution is a generalized extreme value (GEV) distribution left truncated at zero (TGEV). The truncation corrects the disadvantage of the GEV distribution-based models occasionally predicting negative wind speed values, without affecting its favourable properties. The new model is tested on four data sets of wind speed ensemble forecasts provided by three different ensemble prediction systems, covering various geographical domains and time periods. The forecast skill of the TGEV EMOS model is compared with the predictive performance of the truncated normal, log-normal and GEV methods and the raw and climatological forecasts as well. The results confirm the favourable properties of the novel TGEV EMOS approach.