Determine Precipitation Rates from Visible and Infrared Satellite Images of Clouds by Pattern Recognition Technique.

摘要

A more advanced cloud pattern analysis algorithm was subsequently developed to take the shape and brightness of the various clouds into account in a manner that is more consistent with the human analyst's perception of GOES cloud imagery. The results of that classification scheme were compared with precipitation probabilities observed from ships of opportunity off the U.S. east coast to derive empirical regressions between cloud types and precipitation probability. The cloud morphology was then quantitatively and objectively used to map precipitation probabilities during two winter months during which severe cold air outbreaks were observed over the northwest Atlantic. Precipitation probabilities associated with various cloud types are summarized. Maps of precipitation probability derived from the cloud morphology analysis program for two months and the precipitation probability derived from thirty years of ship observation were observed.