A great deal of essential information is lost in the widely used quick-look pictures covering large areas and produced in black and white for visible and infrared radiation. The goal of this study was to find an effectiv way to use cloud information in satellite imagery. Through the interactiv image analysis system at the Swedish Space Corporation it has been possible to make multispectral classification of AVHRR data. This classification of different cloud types provides the forecaster with a cloud image on a colour video display. Different stages of convection in different colours indicates stability and moisture conditions in the troposphere.
A qualitative measure of the precipitation intensity has been derived from the satellite imagery of nimbostratus and cumulonimbus.
The classification also implies a very great reduction of the data volume in the satellite imagery . This, in turn makes it possible for remotely situated weather services with telephone connextion to use processed satellite information.