2016 Ocean Sciences Meeting

February 21-26

Data assimilation allows us to combine numerical models and observations in a quantitative way. Both, models and observations are relevant information sources about the state of a system. However, models can only represent the processes that are implemented in their code and model predictions depend on the proper initialization of the model state estimate. On the other hand, oceanic observations of physical and biogeochemical variables represent the reality, but are quite limited. In particular, satellite data only represents the ocean close to the surface and in situ measurements are very sparse. With data assimilation one can obtain a joint estimate of the state, but also estimates of unobservable fluxes and of parameters controlling model processes. The tutorial will provide an introduction to data assimilation in general and with some focus on assimilation methods that use ensembles. Ensemble methods estimate the model uncertainty using an ensemble of model realizations. They can be implemented with very small changes to a numerical model and use today's parallel computers in a very efficient way. The tutorial should be useful for scientists to get an overview of how data assimilation can be performed and what can be achieved using data assimilation.

• B - Biogeochemistry and Nutrients
• OD - Ocean Observing and Data Management
• PC - Past, Present and Future Climate
• PO - Physical Oceanography/Ocean Circulation