Optimal filtering and prediction of observations in oceanography on the basis of Kalman filters

Kalman´s method of filtering developed in the theory of control is highly promising for the application to oceanography. This can be attributed to two reasons: (i) It permits one to take into account the physical properties of the predicted process in the development of an algorithm for prediction of its values; (ii) It allows one to perform filtering and prediction of the process on the electronic computer, using a short series of observations. Let us discuss a time series of oceanographic observations with the normal distribution of probabilities of values. The problem of prediction of the process in the ocean, given by this series of observations, can be solved effectively by the method of the space of states which is well-known in the theory of control. This problem is equivalent to the construction of the optimal linear dynamic system of control.