The General Data Assimilation Method, its Comparison with the Standard Scheme, and its Application to Dynamical Simulation in the Atlantic

Abstract

A new data assimilation scheme developed earlier and based on the theory of diffusion stochastic processes and parabolic differential equations is presented and tested. This scheme is applied to the Hybrid Circulation Ocean Model (HYCOM) and altimetry data base Archiving, Validating and Interpolating Satellite Oceanography Data (AVISO) over the Atlantic. Several numerical experiments are conducted and their results are analyzed. It is shown that the method really assimilates data, makes the output oceanic fields closer to observations and, on the other hand, conserves the model integrals and balance. The tested method is also compared with the Ensemble Optimal Interpolation scheme (EnOI) as a counterpart of the standard Kalman filter method and it is shown that the proposed general method has several advantages, in particular, it provides a better forecast and requires less computational consumptions.