This effort will address data assimilation in models of the ocean circulation. The PI will consider two classes of problems: The use of inverse methods in non-linear quasigeostrophic models to understand quasi-Guassian processes, and The development of front tracking to understand intermittent circulation phenomena, such as strong boundary current flows. The use of data input to constrain both types of models will be studied by generalized inverse procedures and iterative methods. Appropriate functional representations will be sought for realistic oceanic fields, including fronts and eddies.