One Two
Sie sind hier: Startseite WEX-MOP Phase 1 Physically Coherent Probabilistic Weather Forecasts

Physically Coherent Probabilistic Weather Forecasts

Roman Shefzik and Tilman Gneiting, Heidelberger Institut fur Theoretische Studien, HITS gGmbH, Thordis Thorarinsdottier, Norwegian Computing Center, Oslo

Realizing the full potential of such an ensemble forecast requires statistical postprocessing of the NWP output, in that model biases and dispersion errors are to be addressed.  While these methods are well developed, they apply to a single weather variable at a single location and a single look-ahead time only, and they are confined to the model grid.  A key challenge then is the extension to the joint treatment of several variables at groups of points and for time intervals. This will be addressed in project parts led by the Heidelberg group. In particular, we will develop the novel approach of ensemble copula coupling (ECC), which provides an elegant way of restoring physical coherence across weather variables, space and time in postprocessed, gridded ensemble forecasts.


Lerch, S. and Thorarinsdottir, T. L. (2013). Comparing nonhomogeneous regression models for probabilistic wind speed forecasting. Tellus A, 65, 21206.

Schefzik, R., Thorarinsdottir, T. L. and Gneiting, T. (2013). Uncertainty quantification in complex simulation models using ensemble copula coupling. Stat Sci, 28, 616-640.

Schefzik, R. (2013). Ensemble copula coupling as a multivariate discrete copula approach. Available at

Schuhen, N., Thorarinsdottir, T. L. and Gneiting, T. (2012). Ensemble model output statistics for wind vectors. Mon Weather Rev, 140, 3204-3219.