Multi-Model Super-Ensemble Ocean Prediction: An Operational Example Using a Kalman Filter in the Adriatic Sea

摘要

Multi-model Super-Ensembles (SE) aim at combining optimally different models. A dynamic Kalman Filter version of this technique was applied to a unique set of in-situ data and operational ocean models during the Dynamics of the Adriatic in Real-Time (DART) field experiment. The technique was shown to significantly improve forecasting skills.