Numerical Modeling of Synoptic Scale Ocean Dynamics.
Final technical rept.,
OREGON STATE UNIV CORVALLIS
Pagination or Media Count:
Research continued on theory and practice of data assimilation. Results of a study of the application of optimal interpolation 0I, the data assimilation method most commonly used in numerical weather prediction, to a regional data set were published. In that study, hydrographic data from the California Current were assimilated into the Harvard quasigeostrophic open ocean model. Good results were obtained. A study of the application of advanced data assimilation methods to simple highly nonlinear systems which exhibit strongly nonlinear behavior such as bimodality and chaos was completed. Most data assimilation methods were derived under assumptions of linearity, and therefore could be expected to fail when applied to systems which exhibit multiple equilibria or chaos. A finite element quasigeostrophic model of the Kuroshio near the coast of Japan was implemented and tested, and found to exhibit multiple stable equilibria in parameter ranges of physical interest. These multiple equilibria correspond to the observed formation and decay of the large meander inshore of the main current off the coast of Honshu. We plan to apply our newly-developed data assimilation methods for nonlinear systems to this model. Theoretical results pertaining to application of adjoint data assimilation methods to regional and large-scale models were obtained. Appearance of high-wavenumber noise which had been noted in a number of published data assimilation studies was found to be the result of omission of necessary constraints in the original formulation of the methods. AN
- Physical and Dynamic Oceanography
- Operations Research