A Comparison of Cloud Cover Statistics from the GLAS Lidar with HIRS

Abstract

Abstract The cloud dataset from the Geoscience Laser Altimeter System (GLAS) lidar on the Ice, Cloud, and Land Elevation Satellite (ICESat) spacecraft is compared to the cloud analysis of the Wisconsin NOAA High Resolution Infrared Radiation Sounder (HIRS) Pathfinder. This is the first global lidar dataset from a spacecraft of extended duration that can be compared to the HIRS climatology. It provides an excellent source of cloud information because it is more sensitive to clouds that are difficult to detect, namely, thin cirrus and small boundary layer clouds. The second GLAS data collection period from 1 October to 16 November 2003 was used for this comparison, and a companion dataset of the same days were analyzed with HIRS. GLAS reported cloud cover of 0.70 while HIRS reported slightly higher cloud cover of 0.75 for this period. The locations where HIRS overreported cloud cover were mainly in the Arctic and Antarctic Oceans and parts of the Tropics. GLAS also confirms that upper-tropospheric clouds (above 6.6 km) cover about 0.33 of the earth, similar to the reports from HIRS data. Generally, the altitude of the cloud tops reported by GLAS is, on average, higher than HIRS by 0.4 to 4.5 km. The largest differences were found in the Tropics, over 4 km, while in midlatitudes average differences ranged from 0.4 to 2 km. Part of this difference in averaged cloud heights comes from GLAS finding more high cloud coverage in the Tropics, 5% on average but >13% in some areas, which weights its cloud top average more toward the high clouds than the HIRS. The diffuse character of the upper parts of high clouds over tropical oceans is also a cause for the difference in reported cloud heights. Statistics on cloud sizes also were computed from GLAS data to estimate the errors in cloud cover reported by HIRS from its 20-km field-of-view (FOV) size. Smaller clouds are very common with one-half of all clouds being <41 km in horizontal size. But, clouds <41 km cover only 5% of the earth. Cloud coverage is dominated by larger clouds with one-half of the coverage coming from clouds >1000 km. GLAS cloud size statistics also show that HIRS possibly overreports some cloud forms by 2%–3%. Looking at groups of GLAS data 21 km long to simulate the HIRS FOV, the authors found that ∼5% are partially filled with cloud. Since HIRS does not account for the part of the FOV without cloud, it will overreport the coverage of these clouds. However, low-altitude and optically thin clouds will not be reported by HIRS if they are so small that they do not affect the upwelling radiation in the HIRS FOV enough to trigger the threshold for cloud detection. These errors are partially offing.