Abstract
A five-year record of upper air data from 145 stations in the northern tropics was used to determine the statistical structure of meteorological fields, applying the “optimal interpolation” method developed by Gandin. Correlation functions were fitted to scatter diagrams of correlation coefficients as a function of distance in the sense of least squares of sets of preconceived suitable functions. An “optimum density” parameter has been derived for each meteorological element by equalization of “errors of interpolation” to random errors of observation.
