Cloud Detection Using Satellite Measurements of Infrared and Visible Radiances for ISCCP

Abstract

The International Satellite Cloud Climatology Project (ISCCP) began in 1983 to collect and analyze weather satellite radiance datasets to produce a new global cloud climatology as part of the World Climate Research Programme. This paper, the first of three, describes the cloud detection part of the ISCCP analysis. Key features of the cloud detection algorithm are 1) use of space and time radiance variation tests over several different space and time domains to account for the global variety of cloudy and clear characteristics, 2) estimation of clear radiance values for every time and place, and 3) use of radiance thresholds that vary with type of surface and climate regime. Design of the detection algorithm was supported by global, multiyear surveys of the statistical behavior of satellite-measured infrared and visible radiances to determine those characteristics that differentiate cloudy and clear scenes and how these characteristics vary among climate regimes. A summary of these statistical results is presented to illustrate how the cloud detection method works in a variety of circumstances. The sensitivity of the results to changing test parameter values is determined to provide a first estimate of the uncertainty of ISCCP cloud amounts. These test results (which exclude polar regions) suggest detection uncertainties of about 10% with possible negative biases of 5% (especially at night).