Retrieval of atmospheric parameters from radiometric measurements using neural networks

Summary form only given, substantially as follows: the authors report on a simulation study carried out using neural networks to invert radiometric data in retrieving atmospheric characteristics. To this purpose they took advantage of the network simulator SNNS developed at the University of Stuttgart. Training and evaluation sets have been constructed starting from the mid-latitude summer standard atmosphere, including humidity and temperature irregularities, ground-based inversions, liquid clouds. The simulated profiles were used as input to Liebe´s microwave propagation model to compute the brightness temperature that would be measured by each channel of a ground-based radiometer