Building a Realistic Simulation of theAtmospheric State in Radiative Transfer

The simulations for the inverse problem of radiative transfer, even if built with a correct Bayesian approach, do not represent the full source of errors present in the experimental data. We point out two categories of errors (atmospheric model errors and non-Gaussian instrumental errors due to the optics and hardware, that are not considered by standard methods. Moreover, we show cases taken from FORUM simulated radiances using an End to End simulator, where se show how the instrument reacts to a non homogeneousneous filed of view.