AbstractThe direct assimilation in 4D‐Var of principal component (PC) scores derived from Infrared Atmospheric Sounding Interferometer (IASI) spectra has recently been demonstrated. To maximize the exploitation of the IASI instrument, a future step is to consider the extension of the PC approach to the extraction of information from the 4.3‐μm CO2‐absorbing region. Shortwave IASI channels are currently underused compared to similar longwave channels because of day‐night variations in data usability due to departures from local thermodynamic equilibrium (LTE). In this paper, we document the introduction of non‐LTE (NLTE) effects in the PC‐based version of the radiative transfer for TIROS operational vertical sounder (RTTOV) fast radiative transfer model (PC‐RTTOV). The inclusion of NLTE effects in PC‐RTTOV has required the development of a parameterized scheme that allows the fast computation of a NLTE correction to LTE radiances. The fast NLTE model is general enough to be applied to any sensor and can be utilized to add a fast and accurate NLTE correction to polychromatic LTE spectra computed by any general radiative transfer model, including RTTOV, which now incorporates the fast NLTE model developed in this study. The accuracy of the NLTE correction is such that daytime and nighttime radiances can be simulated to almost the same degree of accuracy. The comparison with IASI observations shows that the fast NLTE model presented here performs significantly better than the fast NLTE model incorporated in the previous version of RTTOV and also that improvements have to be made to the simulation of NLTE effects at winter high latitudes.