[Mesa-users] Help with reproducing SN 1999em

[Jared Goldberg]

Hi Soham,

The 99em19 model was made with an older version of MESA and won't
give exactly because many other defaults and under-the-hood choices have
been changed in addition to what's changed in the inlists directly. If you
need the exact model, you will be better served using something between
mesa-r10108 and r10398, closer to 10108, or going through a litany of
changelogs to find out which physics and defaults will change. The exact
model also is on the Zenodo repository for MESA IV
<https://zenodo.org/records/2596697>.  In general, I advise you look
through those inlists and run_star_extras to see what was actually done,
rather than relying on the test suite which has been modified from revision
to revision and is a code unit test suite, and very much NOT a paper
reproducibility suite.

If you're trying to get something closer to that model, however, one thing
that will immediately help will be to raise the value of the convective
mixing length you're using.
The test_suite uses an "other_alpha_mlt" scheme which has a different value
of mixing_length_alpha in the H-rich envelope and the H-poor core. For Red
Supergiants / fully convective envelopes, lower values of alpha yield
larger radii and lower Teff. I believe MESA IV 99em19 model uses alpha=3 in
the envelope, whereas you're using 1.5. This gives systematically large
stellar radii, so to shrink the radius, increase the value of alpha_MLT in
the envelope. The Dutch winds are also affected, even though not directly,
because the effective temperature of the star changes.

*You can change this in your inlists by setting*
*      x_ctrl(21) = 3.0 ! alpha_H*

As for overshooting (1), you can turn off overshooting by setting
*min_overshoot_q
>= 1* , which allows overshooting only when m > Mstar (impossible). See documentation
here
<https://docs.mesastar.org/en/latest/reference/controls.html#min-overshoot-q>
for
that control. You can also set f_ov = 1d-15, or comment out everywhere
where you specify overshooting schemes to begin with. Also, you should use
the exponential overshooting scheme to better approximate MESA IV rather
than step overshooting.


Cheers,
~Jared

On Fri, Mar 28, 2025 at 4:52 PM Mandal, Soham (uqa7ur) via Mesa-users <
mesa-users at lists.mesastar.org> wrote:

> Hello!
>
> I am trying to reproduce the SN 1999em from Paxton et al. 2018, Table 3. I
> am using the test_suites make_pre_ccsn_IIp and ccsn_IIp from MESA-r15140
> (MESA SDK 21.4.1) for this test. I am facing two issues  with this.
>
>
>    1. I can't set the overshoot parameter (f_ov) to zero, even with the
>    step overshoot scheme (MESA says f has to be greater than zero).
>    2. At the time of explosion (at the end of the pre_ccsn run), my model
>    has a T_eff of 3282 K, and R_exp of 930 R_sun, as opposed to 4490 K and 603
>    R_sun from Table 3 of Paxton et al. 2018.
>
>
> May I please ask for suggestions as to how to fix this? I have attached a
> copy of the inlists I use for the pre_ccsn run. Any suggestions will be
> greatly appreciated!
>
> Thanks,
> Soham
>
>
> ------------------------------------------------
> Dr. Soham Mandal
> VITA postdoctoral research fellow
> Department of Astronomy
> University of Virginia
> _______________________________________________
> mesa-users at lists.mesastar.org
> https://lists.mesastar.org/mailman/listinfo/mesa-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20250328/f4438891/attachment.htm>