Cloud detection with SAGE observations

Chuntao Liu, University of Wyoming

The Stratospheric Aerosol and Gas Experiment (SAGE) instrument onboard the ERBS satellite measures the attenuation of solar radiation through the atmosphere during sunrise and sunset events. These multi-wavelength measurements contain information about gases, aerosols and clouds in the atmosphere.

To diagnose the presence of clouds, the algorithm devised by Kent (1983) compares measured atmospheric extinction at two wavelengths. This algorithm was used to generate global and temporal distributions of upper tropospheric clouds. However, due to the assumption of horizontal homogeneity inherent to this method, these results have restricted applicability.

To study how the SAGE observations might respond to clouds of different types and shapes, and to thereby better understand the SAGE measurements of cloud events, the ray-tracing model of Rault (1997) was applied to different cloud situations. Simulations were performed for 3600 different scenarios. The results provide a library of extinction and time profiles, which can be used to provide interpretations of observed SAGE II transmittance profiles. Furthermore, some general features were identified and these provide added insight into how to extract cloud information from the SAGE observations. Due to multitude of possible cloud forms and compositions, a rigorous inversion of observations to specific cloud configuration is not feasible. The aim is to narrow the range of possible scenarios that are consistent with the observations. Methods will be devised to carry out this selection.

The modeling results also indicate the degree of information loss that accompanies routine production of archived data, which contains only 0.5-km averages in the vertical. Thus, a new cloud detection method is needed to fully utilize the SAGE measurements.