Joint probabilistic forecasting of wind speed and temperature using Bayesian model averaging

Ensembles of forecasts are typically employed to account for the forecast uncertainties inherent in predictions of future weather states. However, biases and dispersion errors often present in forecast ensembles require statistical post-processing. Univariate post-processing models such as Bayesian model averaging (BMA) have been successfully applied for various weather quantities. Nonetheless, BMA and many other standard post-processing procedures are designed for a single weather variable, thus ignoring possible dependencies among weather quantities. In line with recently upcoming research to develop multivariate post-processing procedures, e.g., BMA for bivariate wind vectors, or flexible procedures applicable for multiple weather quantities of different types, a bivariate BMA model for joint calibration of wind speed and temperature forecasts is proposed based on the bivariate truncated normal distribution. It extends the univariate truncated normal BMA model designed for post-processing ensemble forecast of wind speed by adding a normally distributed temperature component with a covariance structure representing the dependency among the two weather quantities. The method is applied to wind speed and temperature forecasts of the eight-member University of Washington mesoscale ensemble and of the eleven-member ALADIN-HUNEPS ensemble of the Hungarian Meteorological Service and its predictive performance is compared to that of the general Gaussian copula method. The results indicate improved calibration of probabilistic and accuracy of point forecasts in comparison to the raw ensemble and the overall performance of this model is able to keep up with that of the Gaussian copula method.