1998 April 12
              ATLAS12 models and spectra for individual stars
     ATLAS12 is an opacity sampling model atmosphere program that I wrote
in 1992 to allow computation of models with individual abundances using 
the line data from CD-ROMs 1 and 15 (Kurucz 1993a;c).  It is described in 
Kurucz (1993b) with figures that show sample calculations.  Castelli and 
Kurucz (1994) present an ATLAS12 model for Vega.  There were many problems
with the code that have delayed distribution.  I am making the current 
version of ATLAS12 available on the Web even though it does not yet have 
all the functionality originally advertised, namely, depth-dependent 
abundances and explicit treatment of isotopes, and it still uses the 
ATLAS7-ATLAS9 equation of state.  ATLAS12 is able to compute the same 
models as ATLAS9 which uses pretabulated opacities, plus models with 
arbitrary abundances. 
     When actually working with ATLAS12 sampled fluxes, Castelli and I 
found that, while quite accurate for predicting the total flux, the fluxes 
are not accurate in intermediate or narrow bandpass intervals because the 
sample size is too small.  I wrote a special stripped version of the 
spectrum synthesis program SYNTHE to generate the surface flux for the 
converged model.  Plots showing sample calculations are given in Kurucz 
(1995), the complete spectrum of Arcturus with contributions by individual 
molecules in Kurucz (1994), and the complete spectrum for Sirius in Kurucz 
(1996).  As most of the line positions are predicted, the computed flux 
spectrum is not realistic when compared to a high resolution observed 
spectrum.  It is possible to select only the subset of lines with accurate 
wavelengths, but many lines will be missing. This difficulty can be removed 
only by improving laboratory spectrum analyses.  When more energy levels 
are known, more line positions can be computed accurately.
     The spectrum is computed typically at resolving power 500000.  It is 
actually computed as 17 intensity spectra spread across the disk of the 
star from center to limb.  Then rotationally broadened flux spectra are 
computed for a number of values of v sin i, still at a resolving power of 
500000, by interpolating and integrating over the disk.  All information 
about individual lines is thrown away.  To get the complete information, 
the calculation must be made with the full SYNTHE program.  The intensity 
spectra from different models can be combined to mimic surface features.  
Of course it is not necessary to compute the whole spectrum, but only the 
part of interest.  Any model can be used; the spectrum calculation is not 
coupled to ATLAS12.  These high resolution spectra can be compared directly 
to high resolution observed spectra or they can be instrumentally broadened 
to compare to low resolution spectra or photometry.  It is now feasible to 
compute grids of color indices by computing sections of spectra for every 
model in a grid.
Castelli, F. and Kurucz, R.L. 1994 Models for Vega, A&A 281, 817-832.
Kurucz, R.L. 1993a Atomic data for opacity calculations. Kurucz CD-ROM No.1.
Kurucz, R.L. 1993b A new opacity-sampling model atmosphere program for 
     arbitrary abundances, in Peculiar versus Normal Phenomena in A-type 
     and Related Stars, M.M. Dworetsky, F. Castelli, and R. Faraggiana, 
     eds., A.S.P. Conference Series vol. 44, 87-97.
Kurucz, R.L. 1993c Diatomic Molecular Data for Opacity Calculations.
     Kurucz CD-ROM No. 15.
Kurucz, R.L. 1994 Computation of opacities for diatomic molecules, in
     Molecules in the Stellar Environment, U.G. Jorgensen, ed., Springer-
     Verlag, Berlin, 282-295.
Kurucz, R.L. 1995 Synthetic template spectra, in  Highlights of Astronomy, 
     Vol. 10, I. Appenzeller, ed., 407-409.
Kurucz, R.L. 1996 Status of the ATLAS12 opacity sampling program and of 
     new programs for Rosseland and for distribution function opacity, 
     in ASP Conf. Series Vol. 108, Model Atmospheres and Stellar Spectra, 
     ed. S. Adelman, F. Kupka, and W.W. Weiss, pp. 160-164.