Sensitivity through Adjoints: The numerical model dynamical equations represent the physics governing ocean circulation. The good comparison between the model and observed synoptic transports indicates that the wind forcing and the model dynamics connecting the wind forcing to the transport are both good. However, the numerical model does not provide an immediate indication of the area over which wind stress is most important to forcing the strait transport. While carefully designed numerical model experiments may be performed to provide this insight, the adjoint of the model gives a more direct answer. The adjoint is a method that provides the derivative of a model output (such as the transport through the strait) with respect to the model inputs (such as the wind forcing). NRL is presently constructing numerical model adjoints to understand ocean dynamics and assimilate measurements into models.

The strait transport sensitivity to wind stress at a time lag of 3 h (Fig. 8) indicates that the southerly wind stress off the east Korean coast is most influential. A slightly less sensitive area for transport through the strait lies directly south of the strait. Counterintuitively, wind stress across the relatively shallow Yellow Sea shelf is not a large contributor to the strait transport. The physical mechanism connecting the regions of wind influence to the strait transport is the propagation of oceanic Kelvin waves. An observer moving with the wave would have the coast on their right-hand side when facing the direction of propagation. Wind-generated Kelvin waves in the Yellow Sea propagate away from the Korea Strait, taking the wind information to the Taiwan Strait; Kelvin waves generated in the Sea of Japan propagate along the Korea coast to the Korea Strait.

Using Knowledge to Build Systems: Knowledge of the dynamics controlling exchanges between the interconnected seas is needed to build accurate monitoring and prediction systems. Evaluation through in situ measurements is crucial for confidence in any system. Understanding the model sensitivity is required to know where efforts must be concentrated to provide the maximum payoff. All these research issues are leading to improved operational capability. NCOM is presently being implemented in a global 1/8° system to provide surface currents and temperatures throughout the world.

Acknowledgments: This work is supported by the NRL 6.1 Dynamical Linkage of the Asian Marginal Seas (LINKS) and the NRL 6.1 Error Propagation on the Continental Shelf (EPIC) projects.

[Sponsored by NRL]

Previous Page