MiniLab 2: Plotting Frequencies


In Minilab 2, we’ll take mode frequencies calculated by GYRE and plot them using PGstar. This will involve first adding the frequencies to MESA’s history output; and then modifying inlist_pgstar to set up the plots. As the very first step, make a copy of your working directory from MiniLab 1 (with all the changes you have made):

$ cp -a bellinger-2022-mini-1 bellinger-2022-mini-2
$ cd bellinger-2022-mini-2

Alternatively, if you were unable to get things working with MiniLab 1, then you can grab a working directory for MiniLab 2 from here.

Adding Frequencies to History Output

The standard approach to adding extra columns to history output is to modify the how_many_extra_history_columns and data_for_extra_history_columns hooks. The former defines how many extra columns we want to add; while the latter specifies the data to put into the columns, together with their associated names. Before we make these modifications, however, we have to solve a logistical problem: how do we access the GYRE results outside of the process_mode callback routine?

The answer is to store these results in ‘module variables’ — effectively, global variables that can be accessed from anywhere within run_star_extras.f90.

Adding Module Variables

Let’s add a couple of module variables to store the frequencies of the radial and dipole modes, and the inertias of the dipole modes. Add the following code at the appropriate place near the top of run_star_extras.f90:

! >>> Insert module variables below

    real(dp), allocatable, save :: frequencies(:,:)
    real(dp), allocatable, save :: inertias(:)

The save attributes ensure that the values of the variables are preserved throughout program execution. Note that we declare the variables as allocatable arrays, and therefore need to allocate them in extras_startup and initialize them in extras_start_step. Therefore, find the following code blocks and make the following additions:

! >>> Insert allocation code below


as well as

! >>> Insert additional code below

       do k = 1, 50
          frequencies(1,k) = 0
          frequencies(2,k) = 0
          inertias(k) = 0
       end do

Note that we are also storing the inertias of the dipole modes. That is because we want the most p-dominated modes, and therefore need to select the modes with the lowest mode inertias. We have allocated space for 100 modes, but in practice we will find an a priori unknown amount of around 70 or so.

Setting Module Variables

Let’s now modify the process_mode callback routine to set these module variables. As a reminder, process_mode is called each time GYRE finds a mode. Therefore, we have to add checks to see which mode has been found — and if it is a radial or dipole mode, update one or the other module variable accordingly.

Within process_mode, all data for the mode being processed are stored in the md object. The existing code that prints radial frequencies to the screen already shows us how to access these data:

! Print out degree, radial order, mode inertia, and frequency
print *, 'Found mode: l, n_p, n_g, E, nu = ', &
    md%l, md%n_p, md%n_g, md%E_norm(), REAL(md%freq('HZ'))

Here, md%n_p is a simple integer variable containing the acoustic radial order, while md%freq(...) is a function that returns the mode frequency in the desired units (in this case, Hertz). The REAL(...) wrapper is required because md%freq(...) returns a complex value, with the real part containing the frequency and the imaginary part containing the growth rate. This also prints the spherical degree md%l, the g-mode radial order md%n_g, and the mode inertia md%E_norm().

With these points in mind, we can store the frequencies by adding the following code to the process_mode subroutine. Note that we will calculate the frequencies in microHertz ('UHZ') and then normalize the frequencies by s% nu_max and s% delta_nu in order to make the plots look nicer.

if (md%n_p >= 1 .and. md%n_p <= 50) then

    ! Print out degree, radial order, mode inertia, and frequency
    print *, 'Found mode: l, n_p, n_g, E, nu = ', &
        md%l, md%n_p, md%n_g, md%E_norm(), REAL(md%freq('HZ'))

    if (md%l == 0) then ! radial modes
        frequencies(md%l+1, md%n_p) = (md%freq('UHZ') - s% nu_max) / s% delta_nu

    else if (inertias(md%n_p) > 0 .and. md%E_norm() > inertias(md%n_p)) then
        write (*,*) 'Skipping mode: inertia higher than already seen'
    else ! non-radial modes

        ! choose the mode with the lowest inertia
        inertias(md%n_p) = md%E_norm()
        frequencies(md%l+1, md%n_p) = (md%freq('UHZ') - s% nu_max) / s% delta_nu

    end if
end if

Notice here that we are only saving the dipole mode with the lowest inertia.

Adding History Columns

We’re now in a position to add two extra columns to history output, in which we’ll store the frequencies we’ve calculated. First, edit how_many_extra_history_columns to set the number of columns:

! >>> Change number of history columns below

       how_many_extra_history_columns = 100

Next, add code to data_for_extra_history_columns to set up the names and values of the two extra columns:

! >>> Insert code to set history column names/values below

       do k = 1, 50
          write (names(k),    '(A,I0)') 'nu_radial_', k
          write (names(k+50), '(A,I0)') 'nu_dipole_', k
       end do

       if (s%x_logical_ctrl(1)) then

          ! save the frequencies of the radial and dipole modes
          do k = 1, 50
              vals(k)    = frequencies(1, k)
              vals(k+50) = frequencies(2, k)
          end do


          ! write out zeros for the 2*50 columns
          do k = 1, 100
              vals(k) = 0
          end do


Note that we check s%x_logical_ctrl(1) before setting the vals array; that way, we avoid copying undefined values if running GYRE has been skipped.

Running the Code

With these changes to run_star_extras.f90, re-compile and re-run the code:

$ ./mk
$ ./rn

The history file written to LOGS/ should now contain extra columns, containing the frequency data. An easy way to check this is to use the less command with the -S (chop long lines) flag:

$ less -S LOGS/

(Use the left/right cursors key to scan through the columns).

Plotting the Frequencies

We’re now in a position to add a PGstar panel to our run, showing how the mode frequencies change as the star evolves. The type of panel we’ll use is called a ‘history panel’, which plots columns from the history file as a function of model number or time.

Open up inlist_pgstar, and add the following code at the bottom:

! >>> Insert additional parameters below

! History panel showing frequencies

Grid1_plot_name(5) = 'History_Panels1'

History_Panels1_num_panels = 1
History_Panels1_title = 'Frequencies'
History_Panels1_xaxis_name = 'model_number'
History_Panels1_max_width = 0

History_Panels1_yaxis_name(1) = 'nu_radial_10'
History_Panels1_other_yaxis_name(1) = 'nu_dipole_9'

History_Panels1_same_yaxis_range(1) = .true.

(Here, the first line indicates where in the existing grid layout to place the history panel; the subsequent lines specify what to plot in the panel).

Now re-run the evolution, and consider the following question:

  • Why do the frequencies move in lockstep, with the dipole mode having a nearly constant offset from the radial mode?

The answer can be found by considering the asymptotic relation, which gives that the frequencies of the modes scale with the large frequency separation delta_nu, the spherical degree, and radial order:

\[\nu_{n,\ell} \simeq \Delta\nu\left( n + \ell/2 + \epsilon \right)\]

where \(\nu_{n,\ell}\) is the frequency of a mode with radial order \(n\) and spherical degree \(\ell\); \(\Delta\nu\) is the large frequency separation, and \(\epsilon\) is a phase term.