Simulation of Stratospheric Aerosol and Gas Experiment (SAGE) II cloud measurements using airborne lidar data

The Stratospheric Aerosol and Gas Experiment (SAGE) II satellite instrument measures the attenuation due to stratospheric aerosols, gases, and clouds along a tangential path through the atmosphere from the Sun to the satellite instrument. Data from SAGE II and its predecessor, SAGE I, have been used to study and develop climatologies for high altitude cloud. Because these instruments measure the total attenuation on a long optical path and the data are inverted under the assumption of spherical homogeneity in the region of the measurement, interpretation in terms of individual cloud characteristics is difficult. Airborne lidar data, taken on approximately 8000 lan of flight path over the tropical Pacific Ocean, have been used to simulate high altitude SAGE II cloud measurements and their inversion. The lidar data set, not necessarily typical for the region flown over, showed the general presence of patchy cloud with a mean horizontal cloud dimension of 20-25 km and a vertical thickness of 0.6-0.8 km. Use of the lidar data to simulate SAGE II measurements produces cloud extinction values similar in magnitude and distribution to those obtained from SAGE II. These simulations also show the existence of three possible error conditions that may occur as a result of cloud inhomogeneities along the viewing path. In the first error condition, the true altitude of a cloud may be higher than that found as a result of the SAGE II inversion, errors of 1 km or greater occurring in just under 40% of the simulations, mainly at the lower cloud altitudes. In the second error condition, the inverted cloud extinction may differ from the volume averaged extinction along the horizontal ray path, the former often being biased slightly low. In the third error condition, the presence of nonuniform or isolated cloud patches can result in an apparent negative inverted extinction value just below the cloud. Such values were observed in about one third of the simulations. The first two of these errors have been statistically quantified for application to SAGE II cloud data.