Radiances and brightness temperatures from three near-infrared/infrared channels that are available on most current and past satellite imagers were used to develop automated algorithms for identifying multilayered cloud systems (cloud overlap) and cirrus clouds at night. The cloud overlap algorithm uses information from the 3.75 micron, 11 micron, and 12 micron regions of the spectrum and the cirrus algorithm uses 3.75 micron and 11 micron channel data. The cloud overlap algorithm was developed assuming that a scene with cloud overlap consists of a semitransparent ice cloud that overlaps a lower cloud composed of liquid water droplets. Cirrus clouds are taken to be high ice clouds with a visible optical depth of 5.0 or less. The algorithms are applied to single satellite pixels that are already assumed to be cloudy based on cloud mask information. The utility of each algorithm was demonstrated on two different Moderate Resolution Imaging Spectroradiometer (MODIS) scenes and the cloud overlap algorithm was validated against millimeter radar-derived cloud boundaries. Overall the results show that both algorithms have the potential to be very useful for nighttime cloud studies.

Date Published: 5 January 2005
PDF: 10 pages
Proc. SPIE 5658, Applications with Weather Satellites II, (5 January 2005); doi: 10.1117/12.577640

Published in SPIE Proceedings Vol. 5658:
Applications with Weather Satellites II
W. Paul Menzel; Toshiki Iwasaki, Editor(s)