2009ApJ...691.1634S

We investigate the ability of both LTE and non-LTE (NLTE) models to fit the near-UV band absolute flux distribution, f_λ(λ), and individual spectral line profiles of three standard stars for which high-quality spectrophotometry and high-resolution spectroscopy are available: The Sun (G2 V), Arcturus (K2 III), and Procyon (F5 IV-V). We investigate (1) the effect of the choice of atomic line list on the ability of NLTE models to fit the near-UV band f_λlevel, (2) the amount of a hypothesized continuous thermal absorption extinction source required to allow NLTE models to fit the observations, and (3) the semiempirical temperature structure, T_kin(log τ₅₀₀₀), required to fit the observations with NLTE models and standard continuous near-UV extinction. We find that all models that are computed with high-quality atomic line lists predict too much flux in the near-UV band for Arcturus, but fit the warmer stars well. The variance among independent measurements of the solar irradiance in the near-UV is sufficiently large that we cannot definitely conclude that models predict too much near-UV flux, in contrast to other recent results. We surmise that the inadequacy of current atmospheric models of K giants in the near-UV band is best addressed by hypothesizing that there is still missing continuous thermal extinction, and that the missing near-UV extinction becomes more important with decreasing effective temperature for spectral classes later than early G, suggesting a molecular origin.