A. Radkevich and S. Kato
CERES (Clouds and the Earth's Radiant Energy System) deploys satellite borne instruments measuring solar-reflected and Earth-emitted radiation. One of the primary goals of CERES is to observe top-of-atmosphere (TOA) budget. In addition, surface radiation budget is estimated using radiative transfer (RT) models and observed atmospheric states and cloud, aerosol and surface optical properties. Evaluation of RT model and inputs is possible by comparing modeled and observed broadband radiances. In addition, an agreement of modeled and observed radiances enhances a confidence level of both surface radiation budget estimated by RT models and CERES instrument calibrations. Similarly, an agreement between modeled irradiance and irradiance derived from CERES observed radiance also enhance the confidence level of the RT model and Angular Distribution Model (ADM). ADM-derived and observed CERES radiances and irradiances are compared in earlier studies (Hudson et al. 2010). They reported modeled radiance over Dome-C is lower by 7 % and 9% for CERES instruments on, respectively, Terra and Aqua.
We revisit this issue reported by Hudson et al (2010) for two reasons. 1) We use active sensors (CALIPSO and CloudSat) that are not available for CERES observations used by Hudson et al. (2010) were performed to screen clouds. This will also tests the cloud screening algorithm for snow and ice surfaces developed by Radkevich et al. (2013). 2) Comparisons of modeled and observed radiance by a CERES instrument on Suomi-NPP have not been performed.
The RT model we use is based on DISORT coupled with the correlated-k method (Kato et al. 1999). We used the same approach for the lower boundary condition as in Hudson et al. (2010) with a minor modification related to modeling surface albedo. In modeling the snow surface albedo we found that using more realistic diameters of ice spheres as well as the phase function calculated with Mie theory result in a better agreement with observed surface albedos reported in earlier studies. The RT model is flexible to use observed atmospheric states. In this work we create atmospheric temperature, water valor, and ozone profiles extracted from GEOS-4 reanalysis data over Dome-C. Additional check (Radkevich et al 2013) for clear sky snow conditions was performed to ensure no cloud condition not only at the point of CERES observation but also in nearby vicinity.
A comparison between modeling and actual observations will be performed for data acquired by the CERES instruments onboard EOS Terra, Aqua and possibly Suomi-NPP if CERES SSF data product for Suomi-NPP is available.
Hudson, S.R., S.G. Warren, and S. Kato, 2010, JGR, vol. 115, D20110
Kato S., T.P. Ackerman, J.H. Mather, E.E. Clothiaux, 1999, JQSRT, vol. 62, 109 - 121.
Radkevich, A., K. Khlopenkov, D. Rutan, S. Kato, 2013, J. Atmos. Oceanic Technol., vol. 30, 557 - 568.