Convective systems can develop in a thermodynamically unstable atmosphere. Such systems may quickly reach high altitudes and can cause severe storms. Meteorologists are thus especially interested to identify such storm potentials when the respective system is still in a preconvective state. A number of stability indices have been defined to describe such situations. Traditionally, these indices are taken from temperature and humidity soundings by radiosondes. As radiosondes are only of very limited temporal and spatial resolution there is a demand for satellite-derived indices. Within the Meteorological Product Extraction Facility (MPEF) of the Meteosat Second Generation (MSG) satellite, the derivation of a number of stability indices is envisaged for the MSG field of view. Currently, two different algorithms have been developed. One method is referred to as the 'Physical Method', which tries to reconstruct a temperature and humidity profile from the measured brightness temperatures in the SEVIRI channels. The other method is a 'Statistical Method' where a neural network is trained with actually radiosonde derived stability indices and corresponding modelled brightness temperatures to find a statistical relationship between all relevant data. Both methods can only be used in clear-sky conditions and thus require a careful scenes analysis preceding the retrieval of instability indices.

Both methods have been integrated in the MPEF prototype environment at Eumetsat and are applied to a number of GOES-8 sounder cases. The comparisons of the results of either method help to assess their respective value.