Continued Development of 4D-Variational Data Assimilation and Adjoint- Based Methods of Sensitivity Analysis and Applications Using ROMS.

摘要

The scientific goals of this research project are: 1. To explore the factors (e.g. uncertainties in initial conditions versus those in surface forcing and boundary conditions) that limit the predictability of the circulation in regional ocean models in a variety of dynamical regimes; 2. To compare two state-of-the-art variational data assimilation strategies (4DVAR and IOM) and gain experience using both in regional ocean models; 3. To develop ensemble prediction techniques for regional ocean models; 4. Demonstrate the utility of the ROMS data assimilation framework in a real-time, sea-going environment for prediction studies in the Intra-Americas Sea (IAS) with particular emphasis in the Caribbean Sea. As such, we will demonstrate, as a proof of concept, the utility of adjoint modeling and 4DVAR data assimilation in a real-time operational setting, at sea.