The influence of NLTE effects in Fe I lines on an inverted atmosphere. II. 6301 Å and 6302 Å lines formed in 3D NLTE

Context. This paper forms the second part of our study of how neglecting non-local thermodynamic equilibrium (NLTE) conditions in the formation of Fe I 6301.5 A and the 6302.5 A lines affects the atmosphere that is obtained by inverting the Stokes profiles of these lines in LTE. The main cause of NLTE effects in these lines is the line opacity deficit that is due to the excess ionisation of Fe I atoms by ultraviolet (UV) photons in the Sun.Aims. In the first paper, these photospheric lines were assumed to have formed in 1D NLTE and the effects of horizontal radiation transfer (RT) were neglected. In the present paper, the iron lines are computed by solving the RT in 3D. We investigate the effect of horizontal RT on the inverted atmosphere and how it can enhance or reduce the errors that are due to neglecting 1D NLTE effects.Methods. The Stokes profiles of the iron lines were computed in LTE, 1D NLTE, and 3D NLTE. They were all inverted using an LTE inversion code. The atmosphere from the inversion of LTE profiles was taken as the reference model. The atmospheres from the inversion of 1D NLTE profiles (testmodel-1D) and 3D NLTE profiles (testmodel-3D) were compared with it. Differences between reference and testmodels were analysed and correspondingly attributed to NLTE and 3D effects.Results. The effects of horizontal RT are evident in regions surrounded by strong horizontal temperature gradients. That is, along the granule boundaries, regions surrounding magnetic elements, and its boundaries with intergranular lanes. In some regions, the 3D effects enhance the 1D NLTE effects, and in some, they weaken these effects. In the small region analysed in this paper, the errors due to neglecting the 3D effects are lower than 5% in temperature. In most of the pixels, the errors are lower than 20% in both velocity and magnetic field strength. These errors also persist when the Stokes profiles are spatially and spectrally degraded to the resolution of the Swedish Solar Telescope (SST) or Daniel K. Inouye Solar Telescope (DKIST).Conclusions. Neglecting horizontal RT introduces errors not only in the derived temperature, but also in other atmospheric parameters. The error sizes depend on the strength of the local horizontal temperature gradients. Compared to the 1D NLTE effect, the 3D effects are more localised in specific regions in the atmosphere and are weaker overall.