Non-homogeneous regression models are widely used to statistically post-process numerical ensemble weather prediction models. Such regression models are capable of forecasting full probability distributions and correct for ensemble errors in the mean and variance. To estimate the corresponding...