Inversion skill for limited-area shelf modeling. Part I: An OSSE case study

The ability to compute circulation based on limited data is addressed. The operational context is a 3-day ADCP survey in the Yellow Sea, offshore from a tidally energetic semi-enclosed bay. Velocity data are inverted to deduce boundary conditions for a limited-area simulation model. Skill in fitting, interpolating, and extrapolating the data is judged. A model-generated, shelf-scale climatology serves as Truth for the sampling and for the skill assessment. The simulation model is 3-D and nonlinear. Iteration with a linearized inverse achieves fast, monotonic convergence. The model successfully reconstructs Truth at the data points, within them, and shoreward from them. Skill degrades seaward from the data. The relative importance of tidal and subtidal forcing is assessed for Eulerian velocity and Lagrangian displacement. These findings place upper bounds on skill expectations, since the context is idealistic.