D How to calculate a NLTE model with TMAP

A pre-requiste to understand the following scripts is basic knowledge of unix commands. Lists of the most important can be found easily in the WWW.
In addition, knowledge of the emacs text editor is required.
The meaning of the input cards (traditional name adopted from the formerly used FORTRAN punch cards) for ATOMS2, SETF2, LTE2, PRO2, LINE1, and LINE1_PROF, can be found in this User’s Guide.

D.1 Basics

First, this is a short explanation of the directory structure:

For our scripts, we use the following definitions / shell variables (typically set in the ~/.alias file):

setenv AA ${HOME}/adaten

setenv BI ${HOME}/bimod

setenv FF ${HOME}/fgrids

setenv JO ${HOME}/jobs

setenv MO ${HOME}/models

D.2 Creation of atomic data files with ATOMS2

• Change into the directory /home/${logname}/jobs/atoms2 and copy the template job file to <name>.job. It needs a file <name> with the atomic data in /home/${logname}/adaten. The name <name> can be chosen to contain all the elements used in the model (e.g. H+He)
• You need to edit both files, test_atmos_lte.job and test_atoms.job (e.g. with emacs). The first one is for LTE2, the second one for PRO2.
• After editing, execute the job files and create output files for both jobs with
  

  <name>_atoms.job

  > <name>_atoms.out and

  <name>_atoms_lte.job

  > <name>_atoms_lte.out .

• Check both output files (with emacs) for error messages.
• The jobs create ASCII files, e.g. /home/${logname}/adaten/<name>.atoms2.
  

D.3 Creation of a frequency grid with SETF2

• edit both files <name>_lte_setf2 and <name>_setf2 in /home/${logname}/jobs/setf2.
  Adjust the temperature and make sure that it refers to <name>_lte.atoms2
• run both jobs and create output files with

  <name>_setf2

  > <name>_setf2.out and

  <name>_lte_setf2

  > <name>_lte_setf2.out .

• it writes the files <name>_<temperature>_lte.setf2 and <name>_<temperature>.setf2
  to /home/${logname}/fgrids
  

D.4 Adaption of the Parameter files

• extract the parameters from the <name>_atoms.out file with
  grep para <name>_atoms.out and from the output file <name>_setf2.out of the frequency grid
• if you want to save the parameters you can write them in a file with
  grep para <name>_atoms.out > object.out
• the parameter files for PRO2 can be found in /home/${logname}/PARAMETER/pro2
• copy the existing parameter files to PARA_<name>.INC, PARA1_<name>.INC,
  PARA3_<name>.INC and adjust them by using the parameters from
  <name>_atoms.out
  NOTICE: the values in the parameter files should never be 0!
• we have two machines with licenses to compile our codes, PRO2 has to be compiled on ait320 and LINE1 on aithp3:

  Examples:
  compile with
  ssh ait320 load pro2 <name> [x64/p64]
  (use x64 only to create a unified executable for 64 bit Intel and AMD processors – it takes twice the time compared to p64 that should be used for most of our 64 bit machines. The compiled executable is copied to /home/${logname}/bimod/pro2_<name>.Linux_[x64]
  or with
  ssh aithp3 load line1_prof <name> [x64/p64]

D.5 Calculation of a start model with LTE2

• edit <name>_lte.job with emacs. Adjust the temperature, log g, and the abundances of the elements used (as normalized number fraction) and check if the used files have the right name
• input files: <name>_lte.atoms2, <name>_temperature_lte.setf2
• start the job and write it into an output file with
  nice +19 <name>_lte.job > <name>_lte.out
• the lte model is written to /home/${logname}/model/<name>/lte. The file is called
  temperature_logg_numberfraction element a_numberfraction element b,
  e.g. 0100000_7.00_1.000_0.000
  

D.6 Calculation of a NLTE model with PRO2

• edit the <name>_nlte.job in emacs. Adjust the temperature, log g and the
  abundances of the used elements and check if the used files have the right name
• input files: <name>.atoms2, <name>_temperature.setf2 and the LTE2 model as input model
• start the job with
  nice +19 <name>_nlte.job > <name>_nlte.out
• the output model is written to /home/${logname}/model/<name>
  

D.7 Strategy to calculate a NLTE model with PRO2

To avoid extremely long calculation times, a reasonable way is to calculate first a model using relatively small model atoms and subsequently, after its convergence, perform a so-called line-formation calculation with much more detailed model atmos on a fixed temperature (and density) structure of the model atmosphere. This yields the NLTE occupation numbers of all considered levels. Figure 1 shows the scheme of this startegy.

Some models calculate without any problem using the LTE2 start model and an atomic data file that contains all lines. This can be tested first. In case that this approach fails, try to follow the steps summarized in Table 5. There is, however, no recipe to guarantee numerical stable models.

D.8 An Example: A model with hydrogen and helium

In the following example, we will use T_eff = 100000 K, log g = 7 and number fractions H=0.9 and He=0.1. The respective parts that have to be changed by a user in the following scripts are marked in red color. Necessary differences in the job files are indicated in blue. Basic H and He model atoms are provided by TMAD ( http://astro.uni-_tuebingen.de/~TMAD). These have to be combined then in one atomic data file in the directory ${AA}/H+He.

D.8.1 ATOMS2

• Copy the job files in /home/${logname}/jobs/atoms2 to H+He_lte.job and H+He.job. They need a file H+He with the atomic data in /home/${logname}/adaten. Edit both files, H+He_lte.job and H+He.job, (with emacs). Make sure you use the atomic data file H+He.
• After editing, execute the job file and create an output file for both jobs and check the output files for error messages.
  

  ---

  Table 6: Example for ATOMS2 jobs in the case of LTE (LTE2, left) and NLTE (PRO2, right) model-atmosphere calculations. ATOMS2 job for an LTE2 model ATOMS2 job a PRO2 model job #!/bin/sh #!/bin/sh set +x;. ${HOME}/.jobstart set +x;. ${HOME}/.jobstart # do not edit the beginning of this file # do not edit the beginning of this file ######################## ######################## ## own job following ## ## own job following ## ######################## ######################## # # cat > options <<eos cat > options <<eos AUTO ION H    I   10  16 -10 AUTO ION H    I   10  16 -10 AUTO ION HE  II   14  32 -14 AUTO ION HE  II   14  32 -14 CBB-AUTO-FILL CBB-AUTO-FILL CBF-AUTO-FILL ( NOOP) CBF-AUTO-FILL CBX-AUTO-FILL CBX-AUTO-FILL RBB-AUTO-FILL ( NONE) .RBB-AUTO-FILL ( NONE) RBF-AUTO-FILL ( NOOP) RBF-AUTO-FILL RDI-AUTO-FILL ( NONE) RDI-AUTO-FILL ( NONE) END OPTIONS END OPTIONS eos eos # # cp ${AA}/H+He PROATOM cp ${AA}/H+He PROATOM expand options PROATOM > SOURCE expand options PROATOM > SOURCE ${BI}/atoms2${sys} ${BI}/atoms2${sys} ls -l ATOMS ls -l ATOMS # # cp ATOMS ${AA}/H+He_lte.atoms2 cp ATOMS ${AA}/H+He.atoms2 # # # do not edit the rest of this file # do not edit the rest of this file ##################### ##################### set +x; ${HOME}/.jobend ${TMPDIR} set +x; ${HOME}/.jobend ${TMPDIR}

  ---

• Run both jobs and create output files for them, e.g., enter
  H+He.job > H+He.out
  on the unix command prompt. The resulting atomic-data files H+He_lte.atoms2 and H+He_0100000.atoms2 are copied to /home/${logname}/fgrids.

D.8.2 SETF2

• Edit both files H+He_lte.setf2 and H+He.setf2 in /home/${logname}/jobs/setf2. Set the temperature in both files to 100000 K and make sure that the jobs files refer to H+He_lte.atoms2 and H+He.atoms2, respectively. In resulting atomic data file for an LTE2 calculation is temperature independent because not line transitions are considered. The respective parts that have to be changed by a user in the following scripts are marked in red color.

  ---

  Table 7: Example for SETF2 jobs in the case of LTE (LTE2, left) and NLTE (PRO2, right) model-atmosphere calculations. SETF2 job for an LTE2 model SETF2 job for a PRO2 model #!/bin/sh #!/bin/sh set +x;. ${HOME}/.jobstart set +x;. ${HOME}/.jobstart # do not edit the beginning of this file # do not edit the beginning of this file ######################## ######################## ## own job following ## ## own job following ## ######################## ######################## # # TTM=0100000 TTM=0100000 # # type=’_lte’ type=’ ’ # # cp ${AA}/H+He${type}.atoms2 ATOMS cp ${AA}/H+He${type}.atoms2 ATOMS # # ${BI}/setf2${sys} << eos ${BI}/setf2${sys} << eos ${TTM} $TTM} PRINT CHECK PRINT CHECK 1.30E+16 1.30E+16 -1 -1 10 10 10 10 eos eos # # if test -s FGRID if test -s FGRID then then cp FGRID ${FF}/H+He${type}.setf2 cp FGRID ${FF}/H+He_${TT}${type}.setf2 else else echo ’no FGRID created’ echo ’no FGRID created’ fi fi # # ##################### ##################### set +x; ${HOME}/.jobend ${TMPDIR} set +x; ${HOME}/.jobend ${TMPDIR}

  ---

• Run both jobs and create output files for them, e.g., enter
   H+He.job > H+He.out
   on the command prompt. The resulting f frequency-grid files H+He_lte.setf2 and H+He_0100000.setf2 are copied to /home/${logname}/fgrids.

D.8.3 Parameter files

• Extract the parameters from the ATOMS2 and SETF2 output files file with
  cd /home/${logname}/jobs/atoms2
  grep para H+He.out
  cd /home/${logname}/jobs/setf2
  grep para H+He_0100000.out
  
• the parameter files for PRO2 can be found in /home/${logname}/PARAMETER/pro2
• copy the already existing parameter files to PARA_H+He.INC, PARA1_H+He.INC, PARA3_H+He.INC and adjust them by using the parameters extracted from the ATOMS2 and SETF2 output files (see above).
  NOTICE: the values in the parameter files should never be equal to 0 and none of the parameters ending with MAD (but NRBBMAD) has to be changed by the user!
• compile with ssh ait320 load pro2 H+He p64. The executable /home/${logname}/bimod/pro2_H+He.Linux_x64 is created.

D.8.4 LTE2

• Edit /home/${logname}/jobs/lte2/H+He.job with emacs. Set T_eff = 100000 K, log g = 7 and number fractions H=0.9 and He=0.1. The input files are H+He_lte.atoms2 and H+He_lte.setf2.
• Start the job (an example is given below) and write its output into a file.
  nice +19 H+He_lte.job > H+He_lte.out The LTE2 model is written to /home/${logname}/model/H+He/lte. The model is named
  0100000_7.00_0.900_0.100.
  

D.8.5 PRO2

• Edit the /home/${logname}/jobs/pro2/H+He.job in emacs. Adjust the temperature, log gand the abundances of the used elements as before and check if the used files have the right names. The input files are H+He.atoms2, H+He_0100000.setf2 and the model from LTE2 is the input model.
• Start the job (an example is given in table below) with
  nice +19 H+He.job > H+He.out
• The output model is written to /home/${logname}/model/H+He/.

D.9 Naming the TMAP models

In the framework of the Virtual Observatory, TMAP models and SEDs that are calculated from them have to follow a general rule. The models’ names have to start with Teff (TTTTTTT) and log g (G.GG), followed by the element abundances in mass fractions. An example is

0100000_7.00_H:__9.0000E-01_HE:_1.0000E-01 .

This is suitable for a small number of elements, where the level name stays relatively short. Thus, the general form for a model name is then

TTTTTTT_G.GG_ABUND_<nnn> ,

where nnn is a three digit integer code. ABUND_nnn corresponds to a file with the same name (located e.g. in the model directory). It has the form

H9.0000E-01<source ...>
HE3.0000E-05<source ...>
C4.0000E-06<source ...>
N2.0000E-06<source ...>
O1.0000E-05<source ...>
NE2.0000E-03<source ...>
SI8.0000E-06<source ...>
FE1.0000E-03<source ...>
NI2.0000E-05<source ...>

Please use one line per element, starting with the TMAP element code. The abundances are in mass fraction. The entry for the source of the abundances is optional and format free. Avoid to use any string that will be erroneously identified as an element code by a UNIX grep command!

In the TMAP jobs, the following has to be inserted to set some shell variables (${moddir} is the model directory, ${abund_old} and ${abund_new} may be used for different abundances).

#
abund_old="ABUND_001"
abund_new="ABUND_001"
#
#
# abundances in mass fraction (see files)
#
# old abundances
oldabund=${moddir}/${abund_old}
#
Hold=‘grep "H" ${oldabund} | awk ’{ print $2 }’‘
HEold=‘grep "HE" ${oldabund} | awk ’{ print $2 }’‘
Cold=‘grep "C" ${oldabund} | awk ’{ print $2 }’‘
Nold=‘grep "N" ${oldabund} | awk ’{ print $2 }’‘
Oold=‘grep "O" ${oldabund} | awk ’{ print $2 }’‘
NEold=‘grep "NE" ${oldabund} | awk ’{ print $2 }’‘
SIold=‘grep "SI" ${oldabund} | awk ’{ print $2 }’‘
FEold=‘grep "FE" ${oldabund} | awk ’{ print $2 }’‘
NIold=‘grep "NI" ${oldabund} | awk ’{ print $2 }’‘
#
#
# new abundances
newabund=${moddir}/${abund_new}
#
Hnew=‘grep "H" ${newabund} | awk ’{ print $2 }’‘
HEnew=‘grep "HE" ${newabund} | awk ’{ print $2 }’‘
Cnew=‘grep "C" ${newabund} | awk ’{ print $2 }’‘
Nnew=‘grep "N" ${newabund} | awk ’{ print $2 }’‘
Onew=‘grep "O" ${newabund} | awk ’{ print $2 }’‘
NEnew=‘grep "NE" ${newabund} | awk ’{ print $2 }’‘
SInew=‘grep "SI" ${newabund} | awk ’{ print $2 }’‘
FEnew=‘grep "FE" ${newabund} | awk ’{ print $2 }’‘
NInew=‘grep "NI" ${newabund} | awk ’{ print $2 }’‘
#

In the input files of e.g. PRO2 respective cards have to use the XXnew variables (XX is the element):

CHANGEABUNDANCEH${Hnew}MASS FRACTION
CHANGEABUNDANCEHE${HEnew}MASS FRACTION
CHANGEABUNDANCEC${Cnew}MASS FRACTION
CHANGEABUNDANCEN${Nnew}MASS FRACTION
CHANGEABUNDANCEO${Onew}MASS FRACTION
CHANGEABUNDANCENE${NEnew}MASS FRACTION
CHANGEABUNDANCESI${SInew}MASS FRACTION
CHANGEABUNDANCEFE${FEnew}MASS FRACTION
CHANGEABUNDANCENI${NInew}MASS FRACTION