NOAA-NESDIS
Regional And Mesoscale Meteorology Team
Daily Satellite Discussion
Friday June 27, 1997
The general problem of cloud height derivation has been approached from two methods: convert the cloud temperature to height with an assumed temperature sounding or geometric means, like stereo. Using cloud shadows is a geometric approach, providing the height of the cloud edges just from a single visible satellite image. This technique proves useful in geographic areas of sparce data, to find heights of clouds and their overshooting tops.

The idea of using the length of shadows of clouds to estimate cloud height has been considered before, by Purdom and Dills, with good accuracy. G.G. Campbell has reconsidered this problem and written a more practical shadow program. By changing the concept: find the cloud which produced the shadow, a much faster and practical method has been derived. This idea has been implemented in RAMSDIS software. Suppose one locates the edge of a cloud shadow and represents it as a vector: S relative to the center of the Earth. Then the cloud edge must occur somewhere along the line connecting S and the location of the sun. The location of the maximum rate of rise of radiance, or maximum gradient method is used to find the cloud.

Fig. 1 (NOAA GOES-9 Channel 1, 0.67 microns)

click picture to enlarge

In the GOES-9 Visible image, multi-layered cloud decks were used as candidates to measure cloud heights relative to a lower cloud in the deck. The height 42 [5] indicates that this cloud is 4.2 kilometers above the cloud below it, with an uncertainty in height based on 1 pixel variations, in this case 0.5 kilometers. The image is centered somewhere in the Pacific ocean around 2 N. latitude and 168 W. longitude, out near the limb of GOES-9.

Questions or comments on this technique can be directed to G.G. Campbell. campbell@cira.colostate.edu

Carol Vaughn


We welcome your comments and discussion at ramsdis@comet.ucar.edu

Previous RAMMT Daily Satellite Discussions