Comparison of non-homogeneous regression models for probabilistic wind speed forecasting

• Verfasst von: Lerch, Sebastian [VerfasserIn]
• Verfasst von: Thorarinsdottir, Thordis [VerfasserIn]
• Titel: Comparison of non-homogeneous regression models for probabilistic wind speed forecasting
• Verf.angabe: by Sebastian Lerch and Thordis L. Thorarinsdottir
• E-Jahr: 2013
• Jahr: 14 Nov 2013
• Umfang: 13 S.
• Fussnoten: Gesehen am 26.01.2022
• Titel Quelle: Enthalten in: Tellus / A
• Ort Quelle: Abingdon : Taylor & Francis, 1983
• Jahr Quelle: 2013
• Band/Heft Quelle: 65(2013), 1, Artikel-ID 21206, Seite 1-13
• ISSN Quelle: 1600-0870
• DOI: doi:10.3402/tellusa.v65i0.21206
• Datenträger: Online-Ressource
• Sprache: eng
• Sach-SW: ensemble post-processing
• Sach-SW: non-homogeneous regression
• Sach-SW: predictive distribution
• Sach-SW: probabilistic forecasting
• Sach-SW: weather forecasting
• Sach-SW: wind speed
• K10plus-PPN: 1787198758

Abstract

In weather forecasting, non-homogeneous regression (NR) is used to statistically post-process forecast ensembles in order to obtain calibrated predictive distributions. For wind speed forecasts, the regression model is given by a truncated normal (TN) distribution, where location and spread derive from the ensemble. This article proposes two alternative approaches which utilise the generalised extreme value (GEV) distribution. A direct alternative to the TN regression is to apply a predictive distribution from the GEV family, while a regime-switching approach based on the median of the forecast ensemble incorporates both distributions. In a case study on daily maximum wind speed over Germany with the forecast ensemble from the European Centre for Medium-Range Weather Forecasts (ECMWF), all three approaches significantly improve the calibration as well as the overall skill of the raw ensemble with the regime-switching approach showing the highest skill in the upper tail.