Probabilistic Forecasting: Ensemble Model Output Statistics using Heteroskedastic Censored Regression

We propose a novel way of statistically post-processing forecast ensembles using heteroskedastic censored (Tobit) regression, where location and spread derive from the ensemble forecast. The resulting ensemble model output statistics (EMOS) method is applied to wind speed forecasts over the Pacific Northwest using the University of Washington Mesoscale Ensemble and to inflation forecasts from the Survey of Professional Forecasters. The statistically post-processed EMOS density forecasts turn out to be calibrated and sharp, and result in substantial improvement over the unprocessed ensembles