Low- and Midlevel Cloud Analysis Using Nighttime Multispectral Imagery

Abstract

A multispectral cloud analysis technique using NOAA-7 Advanced Very High Resolution Radiometer (AVHRR) infrared imagery was developed and tested using the AFGL Man-computer Interactive Data Access System (McIDAS) and the AFGL Interactive Meteorological System (AIMS). Fractional cloud amount and cloud top heights are computed for low-level clouds at night, including subpixel resolution clouds (i.e., clouds which only partially fill a sensor's field of view). Multispectral analysis offers a technique for detecting low cloud, which is better than cloud analysis using single channel infrared imagery. Theoretical radiances are computed at the 3.7, 10.7 and 11.8 ?m infrared spectral bands of the AVHRR as a function of cloud top altitude and cloud amount for a range of cloud conditions. Satellite-measured radiances are then compared to the theoretical values at each wavelength to determine the best cloud height/cloud amount match for a pixel. Test case comparisons using manually selected clear and partially cloud-filled regions of AVHRR imagery as displayed on AIMS showed good agreement between the multispectral analysis results and evaluation by human interpretation of the images, surface cloud observations and upper air soundings.