A STATISTICAL MODEL OF THE RETRIEVAL OF A HIGH-FREQUENCY COMPONENT OF THE 500-hPa HEIGHT FIELD

A. I. Chavro and E. V. Dmitriev

A statistical model of the inverse problem of the retrieval of a high-frequency component of the 500-hPa height field spectrum in individual regions is constructed by a low-frequency component ( λ > 2500 km) of the global field. It is shown that this problem can successfully be solved with a 45% accuracy when handling the inverse problem in a space of EOF expansion coefficients with a proposed method and using an optimal series summation. Given an accurate evaluation of the reduction operator R, a mean-square error is 27%. High correlation is obtained between the real and retrieved fields (correlation coefficients ≥ 0.70). A method of optimal summation of the Fourier series is proposed for the case of insufficient observation series.