Analysis Models for the Estimation of Oceanic Fields

View More View Less
  • 1 Center for Earth and Planetary Physics, Division of applied Sciences, Harvard University, Cambridge, MA 02138
© Get Permissions
Full access

Abstract

A general model for statistically optimal estimates is presented for dealing with scalar, vector and multivariale datasets. The method deals with anisotropic fields and treats space and time dependence equivalently. Problems addressed include the analysis, or the production of synoptic lime series of regularly gridded fields from irregular and gappy datasets, and the estimate of fields by compositing observations from several different instruments and sampling schemes. Technical issues are discussed, including the convergence of statistical estimates, the choice of representation of the correlations, the influential domain of an observation, and the efficiency of numerical computations.

Abstract

A general model for statistically optimal estimates is presented for dealing with scalar, vector and multivariale datasets. The method deals with anisotropic fields and treats space and time dependence equivalently. Problems addressed include the analysis, or the production of synoptic lime series of regularly gridded fields from irregular and gappy datasets, and the estimate of fields by compositing observations from several different instruments and sampling schemes. Technical issues are discussed, including the convergence of statistical estimates, the choice of representation of the correlations, the influential domain of an observation, and the efficiency of numerical computations.

Save