Method based on a radiative transfer model to extract the solar component from NOAA-AVHRR channel 3

The AVHRR channel 3 radiance of the NOAA satellites has been applied in recent years to a variety of remote sensing applications. In particular, clouds have been the object of numerous studies such as detection, radiative transfer process and classification. Usually, we must extract the two contributions present in this spectral band, solar and thermal radiances. This separation presents some difficulties, mainly in images with cloudy areas. In this paper, we propose a simple method to obtain the reflected solar radiance based on the shape of T3 T4 versus T4 histograms. These scattergrams emphasize the different spectral behaviour of clouds in those bands corresponding to AVHRR channels 3 and 4. To obtain the individual contributions, we analyse first the simple case where solar radiation is not present, that is, night imagery. Then, we propose a simple theoretical model for clouds that fits the satellite data and will be applied to the cloud parametrization in daylight images. Finally, this model will be inverted to obtain both the solar and thermal radiative contributions.