Generative machine learning methods for multivariate ensemble post-processing

Ensemble weather forecasts based on multiple runs of numerical weather prediction models typically show systematic errors and require post-processing to obtain reliable forecasts. Accurately modeling multivariate dependencies is crucial in many practical applications, and various approaches to multivariate post-processing have been proposed where ensemble predictions are first post-processed separately in each margin and multivariate dependencies are then restored via copulas. These two-step methods share common key limitations, in particular the difficulty to include additional predictors in modeling the dependencies. We …

applications cs.lg dependencies ensemble errors generative machine machine learning modeling multiple multivariate numerical numerical weather prediction physics.ao-ph post-processing practical prediction prediction models predictions processing show stat.me weather weather prediction