SNAP Data Processors - OLCI O2A Harmonisation Algorithm Specification |
Due to the measurement principle of OLCI, the central wavelengths of all bands vary over the field of view and
between the camera interfaces. The absolute spread is in the order of 1.3 nm. This is commonly (although
incorrectly) called smile effect. For all bands, not influenced by strong spectral features of the atmosphere or
the surface, the impact on L2 algorithms is very small and additionally, can be corrected by a smile correction.
The smile correction is a preprocessor, working on top of atmosphere reflectance (eliminating the influence of
spectral variations of the solar constant). A smile correction as implemented in the OLCI ground segment processor
is based on a first order Taylor expansion, approximating the spectral sensitivity with a simple difference
quotient of neighboring bands and modifying the bands to be consistent with their nominal wavelength.
For the OLCI bands 13, 14 and 15 (the new bands in the O2A absorption line (764.4 and 767.5 nm) and the
existing channel at 761.25nm) such a standard smile correction is not possible, since
the spectral sensitivity cannot be approximated using neighborhood bands. This is undesirable, since concurrently
the smile impact of OLCI is largest at these bands. Eventually, the smile effect impedes a simple usage of the
oxygen bands for cloud detection.
R. Preusker, Freie Universität Berlin, has developed an easy to use procedure, which modifies the effective
transmittance in bands 13, 14 and 15 in a sense of a 'harmonisation', which means that these 'harmonised'
transmittances are the ones which would be measured at their corresponding mean wavelengths and with
nominal bandwidth.
The algorithm and its underlying equations are described in more detail in the ATBD of the ESA SEOM S34SCI Land
Study for Snow [1].
[1]
Kokhanovsky, A.A., Box, J.E., Lamarre, M., Dumont, M., Picard, G., Danne, O., and C. Brockmann:
Snow Properties Retrieval from Sentinel-3 (Algorithm Theoretical Basis Document).
ESA AO/1-8502/16/I-NB; SEOM S34SCI Land, Study 1: SNOW. May 2018.