Target Metadata: Coordinates, Priors, and Photometry

Target Metadata: Coordinates, Priors, and Photometry#

The built-in Target class gives us a convenient infrastructure to change or view the metadata specific to our target in the simulation.

There are four main attribute categories that can be changed, and each may contain additional modifiable attributes.

The most basic way the Target class can be used is by initializing it with a name.

from stelpar import Target

target = Target('IRAS 04171+2756')

Now, target has been redifined as a Target object instead of just the target name string. It can be used in the estimation in exactly the same way as before, i.e., est = Estimate(target) (see, e.g., the Quickstart Tutorial). Since for now we only set the Target object with a target name, it behaves the same within the simulation as setting target = 'IRAS 04171+2756'.

Using Coordinates#

It’s possible to pass coordinates for your target if you don’t have a name that’s recognizable by the databases. When you initialize a Target object, you can pass either an astropy.coordinates.SkyCoord object, or you can pass the coordinates as a tuple and any keyword arguments that you would use to create a SkyCoord object. See the astropy documentation for more details.

We still need to give it a name, but now we can call it whatever we want since the name won’t be used to query the photometry.

Let’s create an arbitrary target called “my target” which we’ll say has a right ascension and declination of 100 degrees and -50 degrees, respectively. Once it’s initialized, we can treat it exactly as any other Target object.

my_target = Target('my target', (100, -50), unit='deg')
my_target.coords

<SkyCoord (ICRS): (ra, dec) in deg
    (100., -50.)>

Setting Initial Conditions#

The second metadata category includes all of the initial/boundary conditions for the simulation. There are four initial condition attributes that we can manipulate: bounds, prior, position, and perturbation.

  • Bounds: the hard boundaries of the simulation, beyond which the log-probability goes to \(-\infty\) (i.e., the answer cannot lie outside the bounds). Each fit parameter (age, mass, \(A_v\), f) has upper and lower bounds that can be easily changed if necessary. Default bounds for age and mass are constrained by the isochronal model that is used by the simulation. Default bounds for \(A_v\) and f are loose physical constraints. Bounds are given as (upper, lower) pairs.

  • Prior: the Gaussian (normal) distributions for any of the fit parameters as a way to constrain them. You can also set a prior for \(T_{\mathrm{eff}}\), which is applied as an additional prior to age and mass since these two parameters are used to interpolate \(T_{\mathrm{eff}}\) from the model grid. Priors are given as (\(\mu\), \(\sigma\)) pairs, where \(\mu\) and \(\sigma\) are the mean and standard deviation of the distribution, respectively. Priors are all user-defined, thus none are applied by default.

  • Position and Perturbation: these attributes combine to form the initial position for each fit parameter of the MCMC simulation. The initial position of each walker (in parameter space) is sampled from a normal distribution centered on the given position and scaled by the perturbation.

For reference, we can view the default initial conditions in target.

target.initial_conditions
bounds prior position perturbation
target parameter
IRAS 04171+2756 age (1.0, 10000.0) (nan, nan) 700 35
mass (0.09, 2.45) (nan, nan) 1 0.2
Av (0.0, 3.0) (nan, nan) 0 0.2
f (0.0, 2.0) (nan, nan) 0 0.2
Teff (nan, nan) (nan, nan) NaN NaN

López-Valdivia et al. 2021 (DOI: 10.3847/1538-4357/ac1a7b) gives \(T_{\mathrm{eff}} = 3407 \pm 198 \, \mathrm{K}\) for this star. We also know this is a star in the Taurus Star Forming Region (1-5 Myr), so let’s say \(\mathrm{age} = 3 \pm 2 \, \mathrm{Myr}\) (we won’t change the bounds here so the walkers explore the parameter space). Let’s put priors on these parameters to constrain the final result of the MCMC. We can also lower the initial position and perturbation of the age, but this isn’t necessary if there are enough steps in the MCMC. Finally, since Taurus is an active star forming region, let’s allow \(A_v\) to be higher than 3 mag. These initial conditions are set using nested dictionaries for each attribute, and after we set them, we can check them again to make sure they’ve been applied.

target.initial_conditions = {
    'prior' : {
        'age':(3, 2),
        'Teff':(3407, 198)
    },
    'position' : {
        'age':20
    },
    'perturbation' : {
        'age':10
    },
    'bounds': {
        'Av':(0, 6)
    }
}

target.initial_conditions
bounds prior position perturbation
target parameter
IRAS 04171+2756 age (1.0, 10000.0) (3, 2) 20 10
mass (0.09, 2.45) (nan, nan) 1 0.2
Av (0, 6) (nan, nan) 0 0.2
f (0.0, 2.0) (nan, nan) 0 0.2
Teff (nan, nan) (3407, 198) NaN NaN

Adding/Removing Photometry#

Adding Photometry

Depending on the target, it might be useful to add or remove photometry. You might prefer a different catalog that contains photometry that you’d like to use. Or you might have your own photometry that you’ve taken that you prefer to use over the existing values. If you find that the photometry is saturated or wrong in one or multiple bands you can remove these as well.

First let’s look at the photometry that stelpar tries to collect for each target by default.

target.photometry_meta
magnitude error system isochrone_analog
catalog band
II/246/out 2mass_jmag Jmag e_Jmag VEGA 2mass_jmag
2mass_hmag Hmag e_Hmag VEGA 2mass_hmag
2mass_kmag Kmag e_Kmag VEGA 2mass_kmag
I/355/gaiadr3 gaia_gmag Gmag e_Gmag VEGA gaia_gmag
gaia_bpmag BPmag e_BPmag VEGA gaia_bpmag
gaia_rpmag RPmag e_RPmag VEGA gaia_rpmag
V/154/sdss16 sdss_gmag gmag e_gmag AB sdss_gmag
sdss_rmag rmag e_rmag AB sdss_rmag
sdss_imag imag e_imag AB sdss_imag
sdss_zmag zmag e_zmag AB sdss_zmag
II/336/apass9 johnson_bmag Bmag e_Bmag VEGA johnson_bmag
johnson_vmag Vmag e_Vmag VEGA johnson_vmag
I/259/tyc2 tycho_bmag BTmag e_BTmag VEGA tycho_bmag
tycho_vmag VTmag e_VTmag VEGA tycho_vmag
I/239/hip_main hipparcos_hpmag Hpmag e_Hpmag VEGA hipparcos_hpmag

Suppose we want to add BT and VT magnitudes from Hipparcos if they’re available (in addition to the ones from TYCHO). We would pass a nested dict to the add_photometry method of our Target instance. The first level is the name of the catalog that is queried. The second level contains the names that we assign to the bands. These names are only relevant in our photometry DataFrame (of the Estimate class) so they can be whatever we want them to be. The important information is contained beyond these band keys in a tuple which contains four strings. The first two are the column names in the queried catalog of the magnitude and error. The third is the magnitude system (either ‘vega’ or ‘AB’). The fourth value tells stelpar which of the bands in the isochronal grid is analogous to this new band. This is an optional requirement, but highly recommended. If no analog is given, the band name is used. The list of available analogs is given by the isochrone_analogs attribute of our Target instance.

target.add_photometry(
    {
        'I/239/hip_main' : {
            'hipparcos_btmag' : ('BTmag', 'e_BTmag', 'vega', 'tycho_bmag'),
            'hipparcos_vtmag' : ('VTmag', 'e_VTmag', 'vega', 'tycho_vmag'),
        }
    }
)

print("Isochrone analogs:\n")
print(target.isochrone_analogs)

Isochrone analogs:

['2mass_hmag', '2mass_jmag', '2mass_kmag', 'cousins_imag', 'cousins_rmag', 'gaia_bpmag', 'gaia_gmag', 'gaia_rpmag', 'hipparcos_hpmag', 'johnson_bmag', 'johnson_vmag', 'ps1_gmag', 'ps1_imag', 'ps1_omag', 'ps1_rmag', 'ps1_wmag', 'ps1_ymag', 'ps1_zmag', 'sdss_gmag', 'sdss_imag', 'sdss_rmag', 'sdss_zmag', 'tycho_bmag', 'tycho_vmag']

Local Photometry

Sometimes we might have our own photometry that we’ve taken that we’d like to use. These can be added in a similar way as before, except the catalog must be 'local', and the magnitude and error values in the tuple must be floats.

Suppose we’ve measured R and I magnitudes to be 14 and 15 mag, respectively, and each with an error of 0.05 mag. Remember we can name these whatever we want as long as they have an appropriate analog in the isochrone model.

target.add_photometry(
    {
        'local' : {
             'my_rmag' : (14, 0.05, 'vega', 'cousins_rmag'),
             'my_imag' : (15, 0.05, 'vega', 'cousins_imag')
        }
    }
)

Removing Photometry

Just as we can add photometry, we can also remove any catalogs/bands from being included in the photometry search. This may be useful if, for example, the photometry is saturated in a particular band or for some reason it’s obvious the photometry is wrong. Removing photometry is simpler than adding since all we’re doing is deleting rows from the above table.

We’ll use the remove_photometry method of our Target object and pass a dictionary whose keys are the catalogs and whose values are either the band, list of bands, or ‘all’. The first two are used for removing an individual band(s) and the third is for removing the whole catalog.

Let’s suppose we want to remove 2MASS H, SDSS g, SDSS z, and both Johnson magnitudes from being collected.

target.remove_photometry(
    {
        'II/246/out' : '2mass_hmag',
        'V/154/sdss16' : ['sdss_gmag', 'sdss_rmag'],
        'II/336/apass9' : 'all'
    }
)

Finally let’s check that all of these changes have been applied.

target.photometry_meta
magnitude error system isochrone_analog
catalog band
II/246/out 2mass_jmag Jmag e_Jmag VEGA 2mass_jmag
2mass_kmag Kmag e_Kmag VEGA 2mass_kmag
I/355/gaiadr3 gaia_gmag Gmag e_Gmag VEGA gaia_gmag
gaia_bpmag BPmag e_BPmag VEGA gaia_bpmag
gaia_rpmag RPmag e_RPmag VEGA gaia_rpmag
V/154/sdss16 sdss_imag imag e_imag AB sdss_imag
sdss_zmag zmag e_zmag AB sdss_zmag
I/259/tyc2 tycho_bmag BTmag e_BTmag VEGA tycho_bmag
tycho_vmag VTmag e_VTmag VEGA tycho_vmag
I/239/hip_main hipparcos_hpmag Hpmag e_Hpmag VEGA hipparcos_hpmag
hipparcos_btmag BTmag e_BTmag vega tycho_bmag
hipparcos_vtmag VTmag e_VTmag vega tycho_vmag
local my_rmag 14 0.05 vega cousins_rmag
my_imag 15 0.05 vega cousins_imag

We can see that BT and VT magnitudes were added from Hipparcos, local R and I photometry were added, and 2MASS H, SDSS g, SDSS z, and both Johnson magnitudes were removed.

Reminder: Adding photometry does not guarantee it will be used in the simulation, only that it will be included in the photometry query. If photometry cannot be found in the default or added catalogs, it will not be used (unless it's given as 'local' photometry).

Misc.#

User-Input Parallax

Sometimes stelpar can’t find a viable parallax and/or error (usually Gaia doesn’t have a parallax measurement). In these cases, you can provide your own. This happens in much the same way as the other metadata discussed above. Regardless of the existence of a Gaia parallax, stelpar will always use the user-input parallax.

Note: Parallax and error must be given in arcsec. 
target.plx = (7.8e-3, 1.9e-5) # (parallax, error) [arcsec]
target.plx

(0.0078, 1.9e-05)
Contents