Context. Solar spectral irradiance (SSI) variability is one of the key inputs to models of the Earth’s climate. Understanding solar irradiance fluctuations also helps to place the Sun among other stars in terms of their brightness variability patterns and to set detectability
limits for terrestrial exoplanets.
Aims. One of the most successful and widely used models of solar irradiance variability is Spectral And Total Irradiance REconstruction model (SATIRE-S). It uses spectra of the magnetic features and surrounding quiet Sun that are computed with the ATLAS9
spectral synthesis code under the assumption of local thermodynamic equilibrium (LTE). SATIRE-S has been at the forefront of solar
variability modelling, but due to the limitations of the LTE approximation its output SSI has to be empirically corrected below 300 nm,
which reduces the physical consistency of its results. This shortcoming is addressed in the present paper.
Methods. We replaced the ATLAS9 spectra of all atmospheric components in SATIRE-S with spectra that were calculated using the
Non-LTE Spectral SYnthesis (NESSY) code. To compute the spectrum of the quiet Sun and faculae, we used the temperature and
density stratification models of the FAL set.
Results. We computed non-LTE contrasts of spots and faculae and combined them with the corresponding fractional disc coverages,
or filling factors, to calculate the total and spectral irradiance variability during solar cycle 24. The filling factors have been derived
from solar full-disc magnetograms and continuum images recorded by the Helioseismic and Magnetic Imager on Solar Dynamics
Observatory (SDO/HMI).
Conclusions. The non-LTE contrasts yield total and spectral solar irradiance variations that are in good agreement with empirically
corrected LTE irradiance calculations. This shows that the empirical correction applied to the SATIRE-S total and spectral solar
irradiance is consistent with results from non-LTE computations.