A New Coupled 4DVAR Assimilation System for Coupled Ocean-Wave Models

The coastal ocean environment poses prediction challenges due to shortened time and space scales and highly nonlinear interactions between its wave and circulation dynamics. In-situ observations in such a region can often be scattered and/or incomplete, further stressing our ability to accurately forecast coastal parameters such as currents, wave heights and direction, and density structure. To extend predictability in these coastal environments, we have developed a coupled, four-dimensional, variational (4DVAR) assimilation system for coupled ocean-wave models. Coupling the assimilation systems for the ocean and waves insures dynamical consistency of the assimilation innovations in a highly nonlinear, continuously evolving environment. Coupled assimilation also permits observations of one process to affect the forecast of another, which can maximize our use of limited datasets.

The ocean-wave assimilation system is presently coupled through three mechanisms: ocean currents, Stokes' drift, and wave radiation stress gradients, and is incorporated within the Coupled Ocean Atmosphere Mesoscale Prediction System (COAMPS). The coupled assimilation is realized by including each of these coupling terms within the adjoint and tangent linear components of the ocean-wave, 4DVAR assimilation system. The ocean-wave assimilation and forecast system is then applied to a series of twin experiments off the southern California coast. These twin experiments are designed to show how information flows from observations assimilated in to either the ocean or wave model through the coupled assimilation to influence both ocean and wave model variables. Inferred corrections from one system to another are demonstrated. As part of the coupled assimilation system, a newly developed formulation for wave error covariances is implemented. Impacts of the wave covariances on predictions from the coupled ocean-wave model system are also assessed.