[mesa-users] Kjeldsen correction in astero module

[Santiago Andres Triana]

Hi all,

Just a quick question: In the subroutine get_kjeldsen_radial_freq_corr
in astero_support.f, lines 756 and 757  (version 7385) it reads:

              l0_freq_corr(i) = l0_freq_corr(i) + &
                  correction_factor*(a_div_r/Qnl)*pow_cr(l0_freq(i)/nu_max,b)

What is the reason the argument of pow_cr has l0_freq(i) instead of
l0_obs(i)? as equation (4) in Kjeldsen08 would seem to imply?

Thanks in advance!

Cheers,

Santiago





On Tue, Feb 17, 2015 at 11:40 PM, Santiago Andres Triana
<repepo at gmail.com> wrote:
> Hi Dennis, mesa users,
>
> I'm a bit confused now... (I am doing some tests with no correction
> for the moment) the value reported by delta_nu_model is the 'cheap'
> one (i.e. obtained with an integral involving the Brunt Vaisala
> frequency) when we are still 'cold'. And then, when we are 'warm'
> delta_nu_model is obtained from the radial modes reported by gyre? It
> seems not to be the case... I don't see the 'drop' anymore (but it
> shows when I do corrections)
>
> I did a single run with no correction and saved the models. Then I ran
> gyre manually on these models
> and calculated delta_nu_model independently. it doesn't match what the
> delta_nu_model says from the history file.
>
> I have the feeling I am going to learn something obvious :) but still
> don't know exactly what that is. Thanks again!
>
> Cheers,
>
> Santiago
>
> On Tue, Feb 17, 2015 at 12:00 AM, Dennis Stello
> <stello at physics.usyd.edu.au> wrote:
>> Hi Santiago,
>>
>> I think what you are experiencing is the known offset between the asymptotic
>> delta_nu and that from the 'average' spacing between overtone modes in the
>> observed frequency range.  We start using the asymptotic delta_nu as a guide
>> for how close we are to a good solution because it is cheap.  When you get
>> close, we start with the radial modes.
>>
>> As you say, increasing the assumed uncertainty in delta_nu is a way to make
>> sure we will find the solution.
>>
>> However, to make is easier to tune the delta_nu value, the next release of
>> MESA will plot the radial modes in the echelle etc as soon as they are
>> calculated, giving you a better feel for if you need to decrease or increase
>> delta_nu (thanks to Bill).  So after a few test runs, you will have a feel
>> for what delta_nu and delta_nu_sigma are most efficient and still robust.
>>
>> Cheers
>> Dennis/
>>
>>
>>
>> On Mon, 16 Feb 2015, Santiago Andres Triana wrote:
>>
>>> Dear Dennis, Kuldeep & MESA users,
>>>
>>> Thanks for the hints! I did look into the simplex_results.data and the
>>> 'a_div_r' column was all zeros. I think the reason is that I was
>>> providing a 'delta_nu' in the inlist slightly above (about 10%) of the
>>> observed delta_nu mean (I was thinking that it was better to use a
>>> value as close as possible to the asymptotic delta_nu there). This
>>> leads to a correction_b <1 and therefore no correction overall.
>>>
>>> Additionally, I had not provided the orders in l0_n_obs(:). It was
>>> bound to fail indeed.
>>>
>>> Now I'm a bit further after providing the radial orders and providing
>>> the correct delta_nu to match the observed mean (or letting the module
>>> calculate it by setting delta_nu=-1). The only thing is that there is
>>> a drop in delta_nu_model as soon as the module starts using the
>>> l0_freqs (instead of using the Brunt-Vaisala frequency) to calculate
>>> it (see the figure). In my earlier experiments this drop was
>>> preventing me from finding good chi2_radial. Now I can tweak
>>> delta_nu_sigma to make the module calculate delta_nu_model from
>>> l0_freqs sooner.
>>>
>>> Hope this make any sense, please let me know otherwise. Again, many
>>> thanks for further ideas!
>>>
>>> Cheers,
>>>
>>> Santiago
>>>
>>> On Mon, Feb 16, 2015 at 10:21 PM, Dennis Stello
>>> <stello at physics.usyd.edu.au> wrote:
>>>>
>>>> Hi Santiago,
>>>>
>>>> It sounds to me that what you are looking for is in the simplex results
>>>> file....did you you look there?
>>>>
>>>> Cheers
>>>> Dennis/
>>>>
>>>> -------
>>>> DENNIS STELLO | Senior Lecturer
>>>> School of Physics A28 | Faculty of Science
>>>>
>>>> THE UNIVERSITY OF SYDNEY
>>>> Rm225, K90 | The University of Sydney | NSW | 2006
>>>> T +61 2 9036 5108  | F +61 2 9351 7726
>>>> E stello at physics.usyd.edu.au  | W http://www.physics.usyd.edu.au/~stello/
>>>>
>>>>
>>>>
>>>> On Sat, 14 Feb 2015, Santiago Andres Triana wrote:
>>>>
>>>>> Hi MESA users,
>>>>>
>>>>> Playing around with the astero module and simplex, I noticed that the
>>>>> frequencies, as reported in the sample_#.data files, are not
>>>>> corrected. They seem to be the raw, uncorrected  frequencies from gyre
>>>>> although I have specified the 'kjeldsen' method in
>>>>> inlist_astero_search_controls.
>>>>>
>>>>> The inlist is attached. The relevant snippet is this:
>>>>>
>>>>>    ! surface corrections
>>>>>      correction_scheme = 'kjeldsen' ! options are
>>>>>         ! 'kjeldsen'   Correction of Kjeldsen et al. (2008)
>>>>>         ! 'cubic'      Cubic correction of Ball & Gizon (2014, eqn 3)
>>>>>         ! 'combined'   Combined correction of Ball & Gizon (2014, eq 4)
>>>>>         ! ''           no corrections
>>>>>      correction_factor = 1
>>>>>         ! use this fraction of the correction; set to 0 to skip doing
>>>>> corrections.
>>>>>      l0_n_obs(:) = -1 ! the observed radial orders (ignored if < 0)
>>>>>         ! the observed radial orders are used in calculating surface
>>>>> corrections
>>>>>         ! if <= 0, use default calculation for radial orders
>>>>>      correction_b = 4.25d0
>>>>>
>>>>>      save_next_best_at_higher_frequency = .true.
>>>>>      save_next_best_at_lower_frequency = .true.
>>>>>
>>>>>
>>>>> I'm convinced that I'm doing something silly somewhere but cannot
>>>>> figure out where... any hints or comments are very much appreciated.
>>>>> Many thanks in advance!
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Santiago
>>>>>
>>>>
>>>
>>