The paper describes a technique for processing multispectral NOAA AVHR
R data to support local oceanographic monitoring. Current fields are c
onstructed using an algorithm based on a method of maximum cross-corre
lation, taking into account the reliability of the correlation and the
incidence of non-oceanic elements in the image. The MCSST method for
multichannel atmospheric correction is compared with a method using te
st areas. The accuracy of estimates of the sea surface temperature was
0.25-degrees-C according to the results of comparison with in situ me
asurements in the coastal upwelling zone.