Forecasting Applications of High-Resolution Satellite Cloud Composite Climatologies

Abstract

In this paper, the authors describe experimental forecasting tools developed from high-resolution satellite cloud composites. The satellite data were extracted from the new 5-km, hourly, global satellite database called Climatological and Historical Analysis of Clouds for Environmental Simulations (CHANCES). Analysis was focused on a region over the former Yugoslavia and Adriatic Sea during summer 1994. Cloud composite images were constructed using digital infrared data for each hour of the day. The value at each pixel in the cloud composites was the fractional coverage of cloud at that location for the season and represented its systematic variation. Composite images were also constructed for conditional probabilities of cloud 1?12 h past each hour of the day. The values at any particular pixel in the composites represented the conditional probability of cloud given an initial condition of cloudy or clear in that pixel. Data from both types of composite images were combined to produce a climatological forecasting tool. Forecast tables were constructed of values for the pixel centered over Sarajevo. These tables are similar to the conditional climatology statistics familiar to forecasters in any weather station. A more sophisticated type of conditional probability was tested in which the initial condition was dependent on the average conditions of a region of pixels surrounding the Sarajevo pixel. Results demonstrate powerful operational applications of high-resolution satellite cloud climatologies.