Modeling of TOA radiance measured by CERES and SCIAMACHY over the East Antarctic Plateau

Abstract:

CERES and SCIAMACHY are satellite borne remote sensing instruments measuring solar-reflected and Earth-emitted radiation at the top-of-atmosphere (TOA). CERES instruments are designed to monitor the Earth's radiation budget by measuring radiation in 3 broad bands. SCIAMACHY sensor measured earth reflected radiation in the spectral range 0.24 to 2.38 um with fine spectral and coarse spatial resolutions. In this work we evaluate CERES shortwave (SW) TOA radiance over permanent clear sky snow in the East Antarctic Plateau to test consistency between modeled and observed radiances. We use SCIAMACHY observations to validate spectral performance of our radiative transfer (RT) model.

We revisiting the issue reported by Hudson et al (2010) with another radiative transfer model and using instantaneous atmospheric profiles. That paper reported some overestimation of TOA albedo by their model in comparison with CERES observed SW radiances. As pointed out by Hudson et al., that comparison involves some uncertainties including errors in the modeled surface albedo and atmospheric properties.

We use RT model based on DISORT coupled with a k-distribution approach (Kato et al 1999). We use the same approach for the lower boundary condition as in Hudson et al. (2010) with a modification related to modeling surface albedo. In this work we create atmospheric profiles for the individual CERES and SCIAMACHY observations from GEOS-4 reanalysis.

A comparison between modeling and actual observations was performed for data from the CERES sensors onboard EOS Terra and Aqua, and Suomi-NPP. Similar to the study by Hudson et al. (2010), the model overestimates the TOA radiance. Modeled radiances are greater than observed ones from the CERES Single Satellite Footprint data by 4.6% for FM-1, 2, and FM-4, and by 3.6% for FM-5. Modeled and observed radiance correlates well: coefficient of determination R² > 0.999.

We compare modeled radiances SCIAMACHY radiances by spectrally integrating over the RT model bands with modeled radiances over the same region of interest in the spectral region 0.25 um to 1.6 um. The difference depends on spectral region. In general, the model overestimates measurements by ~2.9%. Sensitivity of the spectral difference to the snow grain size used in the snowpack model and to water vapor and ozone amounts will be presented.