[mesa-users] Understanding mix_info

[Warrick Ball]

Ehsan,

The way I understand this part is that mt1 and mt2 are simply sorted so 
that the subsequent comparisons, which determine what kind of boundary is 
present, can be made knowing mt1 < mt2.  If this part wasn't present, I'm 
pretty sure all that would happen is that you'd end up with a set of 
equivalent comparisons for the mt1 > mt2 cases.

The interpolations (lines using find0()) are linear and thus symmetric. 
So the code isn't really looking in either direction, inward or outward, 
its just finding convective boundaries by interpolating linearly between 
points where grad_r-grad_a changes sign.  This corresponds to the second 
misuse described in the recent preprint by Gabriel et al. [1].

I hope that makes sense.  I'm trying to take what few spare minutes I can 
muster to look at this problem too.

Cheers,
Warrick

[1] http://arxiv.org/abs/1405.0128



On Mon, 12 May 2014, Ehsan Moravveji wrote:

> Hello all,
> I am trying to better understand how the boundaries of different mixing zones are set by spying into /star/private/mix_info.f. The subroutine set_mlt_boundary_info() starts from the surface and walks inward looking for the zones where there is a transition in mixing type:
>
>         do k = 2, nz
>            mt1 = s% mixing_type(k-1)
>            mt2 = s% mixing_type(k)
>            if (mt1 == mt2) cycle
>
> However, there is suddenly a swap of the definition of the inner and outer layers if the outer has higher-valued mixing type:
>
>            if (mt1 > mt2) then
>               mt = mt1; mt1 = mt2; mt2 = mt
>            end if
>
> What I understand is that in such cases, then we are "locally" looking from inside towards outside.
> If this is the true, please help me understand why, and if I am wrong, clarify why we need such.
>
> Thanks for your attention
> Best regards,
>
> Ehsan
>
>
>
>
>
>


------------
Warrick Ball
Postdoc, Institut für Astrophysik Göttingen
wball at astro.physik.uni-goettingen.de
+49 (0) 551 39 5069