Original paper

Application of numerical weather prediction in wind power forecasting: Assessment of the diurnal cycle

Heppelmann, Tobias; Steiner, Andrea; Vogt, Stephan

Meteorologische Zeitschrift Vol. 26 No. 3 (2017), p. 319 - 331

48 references

published: Jun 14, 2017
published online: Mar 10, 2017
manuscript accepted: Dec 20, 2016
manuscript revision received: Dec 14, 2016
manuscript revision requested: Nov 4, 2016
manuscript received: Aug 15, 2016

DOI: 10.1127/metz/2017/0820

BibTeX file


Open Access (paper can be downloaded for free)

Download paper for free


For a secure integration of weather dependent renewable energies in Germany's mixed power supply, precise forecasts of expected wind power are indispensable. These in turn are heavily dependent on numerical weather prediction (NWP). With this relevant area of application, NWP models need to be evaluated concerning new variables such as wind speed at hub heights of wind power plants. This article presents verification results of the deterministic NWP forecasts of the global ICON model, its ICON-EU nest, the COSMO-EU, and the COSMO-DE as well as of the ensemble prediction system COSMO-DE-EPS of the German National Weather Service (DWD), against wind mast observations. The focus is on the diurnal cycle in the Planetary Boundary Layer as wind power forecasts for Germany exhibit pronounced systematic amplitude and phase errors in the morning and evening hours. NWP forecasts with lead times up to 48 hours are examined. All considered NWP models reveal shortcomings concerning the representation of the diurnal cycle. Especially in summertime at onshore locations, when Low-Level Jets form, nocturnal wind speeds at hub height are underestimated. In the COSMO model, stable conditions are not sufficiently reflected in the first part of the night and the vertical mixing after sunrise establishes too late. The verification results of the COSMO-DE-EPS confirm the deficiencies of the deterministic forecasts. The deficiencies are present in all ensemble members and thus indicate potential for improvement not only in the model physics parameterization but also concerning the physical ensemble perturbations.


wind energyNWPensemble forecastverificationdiurnal cycle of wind speedlow-level jet