Extended Abstract (440K)P3.17
Retrieving surface downwelling longwave radiation from satellite measurements: Improvement of the Zhou-Cess algorithm
Yaping Zhou, Univ. of Maryland Baltimore County, Greenbelt, MD; and D. P. Kratz, A. C. Wilber, S. K. Gupta, and R. D. Cess
Zhou and Cess (2001) based their original algorithm for retrieving surface downwelling longwave radiation (SDLW) upon detailed studies using radiative transfer model calculations and surface radiometric measurements. Their algorithm linked clear sky SDLW with surface upwelling longwave flux and column precipitable water vapor. For cloudy sky cases, they used cloud liquid water path as an additional parameter to account for the effects of clouds. Despite its simplicity, their original algorithm performed very well for most geographical regions. When the algorithm was applied to global conditions, however, large biases were found for certain regions, most notably the Polar Regions where the atmosphere is extremely cold and dry. Systematic errors were also found for regions covered with ice cloud. An improved version of the algorithm prevents the large errors in the SDLW at low water vapor conditions by adding an offset to the logarithmic water vapor term. The new algorithm also utilizes cloud fraction and cloud liquid and ice water paths available from the Cloud and the Earth's Radiant Energy System (CERES) single scanner footprint (SSF) product to separately compute the clear and cloud portions of the flux. The new algorithm has been validated against surface measurements at 29 stations around the globe over a period of 58 months. The results show significant improvement over the original version and are now comparable to more sophisticated algorithms currently implemented in the CERES processing. This revised version of Zhou-Cess algorithm will be incorporated into the CERES operational processing.
Poster Session 3, Radiation Poster Session III: Earth Radiation Budget
Wednesday, 12 July 2006, 5:00 PM-7:00 PM, Grand Terrace
Previous paper Next paper