Objective analysis and classification of oceanographic data

A new approach to the analysis and classification of oceanographic data is presented. The technique is an empirical one, based on the singular decomposition theorem, for characterizing temperature-salinity-depth profiles as in water-mass analysis. Complementary to the profiles are station or cast-dependent coefficients, which show similarities and differences according to the space-time distributions of the individual stations in the data set. These coefficients are used for classifying the individual stations into groups having similar profiles. Results of applying the new method to a set of data from ocean stations off the coast of Oregon are given.