The impact of aerosols on the thermal infrared radiation can be assessed by combining observations and radiative transfer calculations. Both have uncertainties, which are discussed in this paper. Observational uncertainties are obtained for two FTIR instruments operated side by side on the ground during the LACE 1998 field campaign. Radiative transfer uncertainties are assessed using a line-by-line model taking into account the uncertainties of the HITRAN 2004 spectroscopic database, uncertainties in the determination of the atmospheric profiles of water vapor and ozone, and differences in the treatment of the water vapor continuum absorption by the CKD 2.4.1 and MT_CKD 1.0 algorithms. The software package OPAC was used to describe the optical properties of aerosols for climate modeling. The corresponding radiative signature is a guideline to the assessment of the uncertainty ranges of observations and models. We found that the detection of aerosols depends strongly on the measurement accuracy of atmospheric profiles of water vapor and ozone and is easier for drier conditions. Within the atmospheric window, only the forcing of downward radiation at the surface by desert aerosol emerges clearly from the uncertainties of modeling and FTIR measurement. Urban and polluted continental aerosols are only partially detectable depending on the wave number and on the atmospheric water vapor amount. Simulations for the space-borne interferometer IASI show that only upward radiation above transported mineral dust aloft emerges out of the uncertainties. The detection of aerosols with weak radiative impact by FTIR instruments like ARIES and OASIS is made difficult by noise as demonstrated by the signal to noise ratio for clean continental aerosols. Altogether, the uncertainties found suggest that it is difficult to detect the optical depths of nonmineral and unpolluted aerosols.