Detecting radiances in the O-2 A band using polarization-sensitive satellite instruments with application to the Global Ozone Monitoring Experiment

Citation
Dm. Stam et al., Detecting radiances in the O-2 A band using polarization-sensitive satellite instruments with application to the Global Ozone Monitoring Experiment, J GEO RES-A, 105(D17), 2000, pp. 22379-22392
Citations number
37
Categorie Soggetti
Earth Sciences
Volume
105
Issue
D17
Year of publication
2000
Pages
22379 - 22392
Database
ISI
SICI code
Abstract
In this paper, we present numerical simulations of the radiance and the deg ree of linear polarization of light reflected by the terrestrial atmosphere in the O-2 A absorption band, around 760 nm. Since the O-2 A band is often used to derive cloud parameters, we included clouds in our model atmospher e. The simulations show that the polarization of the reflected light change s across the O-2 A band, and that this change depends strongly on the spect ral resolution of the instrument. The polarization of reflected light induc es errors in radiances derived from observations by polarization sensitive instruments. For the Global Ozone Monitoring Experiment (GOME) satellite in strument, which measures radiances with about 0.4-nm spectral resolution in the O-2 A band, broadband polarization measurements are used to correct th e narrowband radiance observations for the instrument's polarization sensit ivity. Although such correction schemes significantly improve the accuracy of derived radiances in the continuum, they do not account for changes of t he polarization in narrow absorption bands, such as the O-2 A band. The mai n purpose of this paper is to investigate for cloudy atmospheres the errors in the derived radiances due to polarization changes across the O-2 A band , both for a polarization sensitive instrument with a high spectral resolut ion and for a GOME-like resolution. If no correction scheme is used, it is found that for nadir viewing directions, the maximum errors in the absorpti on band can increase by up to about 20% with decreasing width of the spectr al response function when the instrument's sensitivity for radiation polari zed perpendicularly to the principal plane is twice as large as that for ra diation polarized parallel to this plane. If, in this case, a correction sc heme based on the broadband value of the polarization is used, the radiance errors can still be up to 18% with a high spectral resolution and of the o rder of a few percent with a GOME-like resolution.