Task 1.3 Hindcast development

In this task, the ICON prediction system will be created, in close collaboration with the ICON-Consortium. The technical handling of the simulations is based on the operational system, e.g. by using the same job schedulers as DWD. The developments and results of Task 1.1 & 1.2 will be – both conceptually and technically – on an iterative basis combined in this task and tested in first assimilation experiments and short hindcast simulations. This work will also draw on the experience from MiKlip that model-consistent assimilation/initialisation is of similar importance as the similarity of model initialisation and observed state (Brune et al., 2018, Brune and Baehr, 2020). The continuous execution and evaluation of assimilation and hindcast simulations is essential to decide which developments from WP1 are ready for implementation in the operational system (with Task 1.5). These simulations will be evaluated with respect to the resulting forecast quality, especially over Europe, also with focus on the signal-to-noise problem, for both the initial operational ICON-XPP system and the refined ICON-XPP system.

Contributors

University of Hamburg
Prof. Dr. Johanna Baehr
Dr. Jürgen Bader
Dr. Sebastian Brune

Max Planck Institute for Meteorology
Dr. Holger Pohlmann

Entire consortium