Retrieval of Cloud Cover Parameters from Multispectral Satellite Images

View More View Less
  • a Goddard Laboratory for Atmospheres Sciences, NASA/Goddard Space Flight Center, Greenbelt, MD 20771
  • | b Systems and Applied Sciences Corporation, Vienna, VA 22180
© Get Permissions
Full access

Abstract

A technique is described for extracting cloud cover parameters from multispectral satellite radiometric measurements. Utilizing three channels (visible, 3.7 μm and 11 μm) from the AVHRR (Advanced Very High Resolution Radiometer) on NOAA polar orbiting satellites, it is shown that one can retrieve four parameters for each pixel: cloud fraction within the FOV, optical thickness cloud-top temperature and a microphysical model parameter. The last parameter is an index representing the properties of the cloud particle (e.g., size, shape, thermodynamic phase, etc.) and is determined primarily by the radiance at 3.7 μm. The other three parameters are extracted from the visible and 11 μm infrared radiances, utilizing the information contained in the two-dimensional scatter plot of the measured radiances. The solution is essentially one in which the distributions of optical thickness and cloud-top temperature are maximally clustered for each region, with cloud fraction for each pixel adjusted to achieve maximal clustering.

Abstract

A technique is described for extracting cloud cover parameters from multispectral satellite radiometric measurements. Utilizing three channels (visible, 3.7 μm and 11 μm) from the AVHRR (Advanced Very High Resolution Radiometer) on NOAA polar orbiting satellites, it is shown that one can retrieve four parameters for each pixel: cloud fraction within the FOV, optical thickness cloud-top temperature and a microphysical model parameter. The last parameter is an index representing the properties of the cloud particle (e.g., size, shape, thermodynamic phase, etc.) and is determined primarily by the radiance at 3.7 μm. The other three parameters are extracted from the visible and 11 μm infrared radiances, utilizing the information contained in the two-dimensional scatter plot of the measured radiances. The solution is essentially one in which the distributions of optical thickness and cloud-top temperature are maximally clustered for each region, with cloud fraction for each pixel adjusted to achieve maximal clustering.

Save