Subject: Re: new opacity files From: David Cohen Date: Mon, 20 Apr 2009 13:36:22 -0400 To: Maurice Leutenegger CC: David Cohen , Erin Martell , Emma Wollman Maurice, We need to revisit our wind opacity models (not that much of what you've sent us here - in this email from February - won't still be useful). But Emma is nearly finished with her individual line profile fitting and she's making good progress on writing up her results in a short paper. And she's at the point where we need to finalize (and produce!) our best-guess universal wind opacity model. We'll need to incorporate this into the windtabs, but actually using it for data modeling is farther in the future (plus, it seems that you've developed a system now whereby we can just swap in new opacity models for use with windtabs). In any case, Emma just needs a tabulation of kappa vs. lambda. Here's what I think Emma's approach should be: - We use a solar metallicity opacity model, but with altered CNO abundances (though C+N+O = solar), as our default wind opacity model. Results (Mdots for individual stars) can then, at any point in the future, be scaled by us or by other researchers as new information about the metallicities of individual stars become available. - We use the fully solar abundance model (kappa1.0.fits here; you've already made it) as a secondary opacity model for Emma to derive mass-loss rates. Same arguments apply to after-the-fact scaling if non-solar metallicities are determined/confirmed for individual stars. Emma can show that the mass-loss rates she derives aren't all that sensitive to the choice of wind opacity models (since they're both solar metallicity, overall, and since there are hardly any lines in any of her stars with wavelengths longward of the O K-shell edge, where the actual CNO abundances matter). But her pairs of Mdot values for each star will nicely bracket the reasonable range of values (again, assuming solar overall metallicity). So... can we ask you to produce such a model for us? It would be identical to kappa1.0.fits but with altered CNO abundances. We have a couple of choices. We can use the relative CNO fractions Janos finds from his cmfgen modeling constrained by the optical/UV spectra. Or we can use Pauldrach's values (I'll dig them up and send them to you soon). Or, we can use some other values that we either make up based on general trends and plausibility arguments or which we find in the literature. What do you think? David Maurice Leutenegger wrote: > Hi all, > > here are new opacity files. Actually, the opacities are not too new, except for the 2/3 solar metallicity model. But they are packaged as FITS files, which may be more convenient to use; if not, it's easy to get text back out of them. > > The files are: > > ZetaPup_Feb_2009_1.7e-6.fits --------------- Janos' most recent model, including all radii; not suitable for windtabs, just more convenient than ascii. > ZetaPup_Feb_2009_1.7e-6_R1.82.fits --------- Same model, with R = 1.82 R_* opacity extracted. Can be used in windtabs. > kappa.66.fits ----------------------------------------- 2/3 solar metallicity > kappa1.0.fits ----------------------------------------- solar metallicity; these last two use AGS2005 abundances, and the same fictional ion balance as before. > > I will send the files out in a second email, since the tar file is ~ 2 MB; you can get the file and then delete that mail. > > I will also send a file called 'tau_transmission.fits' for use with windtabs. None of the ascii text files are used anymore. > > Tonight I will update windtabs on sourceforge. Here are the instructions to update local models and start using windtabs: > > 1. Get the code: > > svn co https://windprofile.svn.sourceforge.net/svnroot/windprofile/trunk/ > > 2. Edit the lmodel.dat entry for windtabs to specify the directory where you will keep the data files for windtabs. The lmodel.dat file that comes with windprofile should already have a sample entry from my installation. The directory must be specified at the end of the first line of the model description. > > 3. There must be two specific files in this directory: tau_transmission.fits, and kappa.fits. kappa.fits may be any of the three files I sent, or more generally any fits file with wavelength and opacity as the first two columns of the first extension. However, you must rename the opacity file to be specifically kappa.fits, unless you want to edit the C++ source code. > > If you want to change between opacity files frequently without recompiling local models, you have two options: a., you can rename a different opacity file to be kappa.fits every time you want to switch, or b., you can keep each opacity file in a different directory (along with a copy of tau_transmission.fits), and then alter lmodel.dat to point to the correct directory whenever you want to change. > > 4. Then you can proceed as usual: initpackage localmodels lmodel.dat (localmodels directory); lmod localmodels (localmodels directory) > > Let me know if you have any questions. > > Maurice