An efficient Markovian algorithm for the analysis of ocean currents

We propose a method for analyzing ocean currents using a statistical approach. The proposed technique is useful for analyzing global velocity fields and producing indices to describe the probable trajectories and destinations of particles embedded in such fields. Short-term Lagrangian integration of the velocities was used to generate transition matrices that define the system locally. A reshuffling algorithm, based on standard Markov Chain theory, was implemented to mix and synthesize the information involved in the global analysis. Iterative methods were then used to solve the resulting large and sparse linear systems. The method efficiently used local information (short-term Lagrangian integration) to infer global characteristics of the system. Two case studies were presented to emphasize the merits of the described scheme: one using modeled data from the Gulf of California, and another from the Gulf of Mexico.