Sergey Frolov
Naval Research Laboratory, Monterey, CA, USA
P. Laloyaux;
M. Bonavita
European Center for Medium Range Weather Prediction, Reading, UK
B. Menetrier
Meteo France, Toulouse, France
The 8th EnKF Data Assimilation Workshop
Implicit and explicit cross-correlations in coupled data assimilation
Poster:
frolov_coupled_da_poster.pdf
Balanced initial condition for the coupled forecast of the ocean and atmosphere requires specification of cross-fluid covariances that can synchronize data assimilation updates in two fluids. If known, for example from an ensemble of coupled forecasts, these covariances can be specified explicitly in a strongly coupled data assimilation (DA) system. Alternatively, these covariances can be generated implicitly by utilizing multiple outer iterations of the four-dimensional DA system. The European Centre for Medium-Range Weather Forecasts has recently developed an implicit coupling approach in the CERA reanalysis system, where otherwise uncoupled atmospheric 4DVAR and ocean 3DVAR are synchronized using 3 outerloop iterations. Since this original work on the outerloop coupling, it has been unclear just how closely does the CERA system with implicit coupled covariances approximate a strongly coupled DA with explicitly specified coupled covariances.
A series of single-observation experiments in a perfect twin framework has been conducted to evaluate the effectiveness of the outer loop coupling and the amplitude of the implicit cross-correlations. They are compared against a coupled assimilation system that explicitly specifies cross-correlations from an ensemble of coupled forecasts. We find that both the implicit outer loop and the explicit ensemble-based methods produce equally skillful estimates of the coupled state in the middle of a 24 hour assimilation window. Our estimates of the speed of the outer loop convergence suggest that atmospheric and ocean states synchronise within the first 6 to 10 hours of the assimilation window. We conclude, that the outer loop coupling is effective when the window is long enough that original imbalances in the atmospheric and ocean increments to synchronise within the length of the assimilation window. On the other hand, we suggest that explicit coupling is preferable for data assimilation systems with short assimilation windows (e.g. 6 hours or less).
Balanced initial condition for the coupled forecast of the ocean and atmosphere requires specification of cross-fluid covariances that can synchronize data assimilation updates in two fluids. If known, for example from an ensemble of coupled forecasts, these covariances can be specified explicitly in a strongly coupled data assimilation (DA) system. Alternatively, these covariances can be generated implicitly by utilizing multiple outer iterations of the four-dimensional DA system. The European Centre for Medium-Range Weather Forecasts has recently developed an implicit coupling approach in the CERA reanalysis system, where otherwise uncoupled atmospheric 4DVAR and ocean 3DVAR are synchronized using 3 outerloop iterations. Since this original work on the outerloop coupling, it has been unclear just how closely does the CERA system with implicit coupled covariances approximate a strongly coupled DA with explicitly specified coupled covariances.
A series of single-observation experiments in a perfect twin framework has been conducted to evaluate the effectiveness of the outer loop coupling and the amplitude of the implicit cross-correlations. They are compared against a coupled assimilation system that explicitly specifies cross-correlations from an ensemble of coupled forecasts. We find that both the implicit outer loop and the explicit ensemble-based methods produce equally skillful estimates of the coupled state in the middle of a 24 hour assimilation window. Our estimates of the speed of the outer loop convergence suggest that atmospheric and ocean states synchronise within the first 6 to 10 hours of the assimilation window. We conclude, that the outer loop coupling is effective when the window is long enough that original imbalances in the atmospheric and ocean increments to synchronise within the length of the assimilation window. On the other hand, we suggest that explicit coupling is preferable for data assimilation systems with short assimilation windows (e.g. 6 hours or less).
Thanks to McGill's Atmospheric and Oceanic Sciences department for hosting this web-site.