Introduced almost 40 years ago, the method of optimal interpolation (OI) ha
s been successfully used in various forms for data analysis in meteorology
and oceanography. At the same time upper-atmospheric applications of OI and
other techniques based on optimal estimation have remained relatively limi
ted. The theory of OI and related methods is reviewed. Properties of optima
l estimation relevant to typical problems of data analysis in the upper atm
osphere are highlighted. As a specific example, a simple one-dimensional sc
heme of OI in time is considered. The scheme is devised to address an impor
tant problem of the upper-atmospheric data analysis: extraction of periodic
(tidal) components from observations covering a fraction of a day. Possibl
e generalizations of the scheme are also briefly discussed. (C) 1999 Elsevi
er Science Ltd. All rights reserved.