An objective technique for Arctic cloud analysis using multispectral AVHRR satellite imagery

Abstract

An established cloud analysis routine has been modified for use in the Arctic. The separation of clouds from the snow and sea ice backgrounds is accomplished through a multispectral technique which utilizes VHRR channel 2 (visible), channel 3 (near infrared) and channel 4 (infrared) data. The primary means of cloud identification is based on a derived channel 3 reflectance image. At this wavelength, a significant contrast exists between liquid clouds and the arctic backgrounds, unlike in the standard visible and infrared images. The channel 3 reflectance is obtained by first using the channel 4 emission temperature to estimate the thermal emission component of the total channel 3 radiance. This thermal emission component is subsequently removed from the total radiance, leaving only the solar reflectance component available for analysis. Since many ice clouds do not exhibit a substantially greater reflectance is channel 3, the routine exploits differences in transmissive characteristics between channels 3 and 4 for identification. The routine was applied to six case studies which had been analyzed by three independent experts to establish 'ground truth'. Verification of the cloud analysis results, through a comparison to the subjective analyses, yielded impressive statistics. A success rate of 77.9% was obtained with an arguably small data base of 131 undisputed scenes