Stellar Kinematics

Analysis class: StellarContinuumModel

Method definition: StellarContinuumModelDef


Reference file: manga-[PLATE]-[IFUDESIGN]-[DRPQA_KEY]-[BIN_KEY]-[CONTINUUM_KEY].fits.gz

Config files: $MANGADAP_DIR/mangadap/config/stellar_continuum_modeling

Example config: gau_mileshc.ini

key                    = MILESHCMPL9
fit_type               = stellar_kinematics
fit_method             = ppxf
fit_iter               = nonzero_templates
reject_boxcar          = 100
minimum_snr            = 1.0
artifact_mask          = BADSKY
emission_line_mask     = ELPMPL8
template_library       = MILESHC
match_resolution       = False
velscale_ratio         = 4
moments                = 2
degree                 = 8
mdegree                = -1

Important class dependencies:

  • StellarKinematicsFit: Provides the abstracted base class for stellar kinematics fits.

  • SpectralPixelMask: Used to mask the fitting regions for the fit.

  • TemplateLibrary: Generalized interface that provides the spectral-template library for use in spectral fitting.

  • PPXFFit: Provides the core pPXF wrapper.


  • Establish the fitting method, which includes setting up the SpectralPixelMask using the artifact_mask, emission_line_mask, and waverange from config.

    • The BADSKY artifact mask is read and used to build an ArtifactDB instance that masks the typical residuals around the strong sky line at 5577 angstroms.

    • The ELPFULL emission-line mask is read and used to build an EmissionLineDB instance that is used to mask the emission lines in the associated parameter file. Each mask is 1500 km/s centered on the line at the input guess redshift.

    • The waverange config parameter can be used to limit the fitted spectral range; will fit as much as possible if no range is provided.

  • Using the SpatiallyBinnedSpectra object, select all binned spectra with S/N greater than minimum_snr in config.

  • The DAP nominally provides the stellar-continuum fit with the velocity and velocity dispersion from The DAP Datacube Configuration File as its initial guess redshift and velocity dispersion.

  • Instantiate the TemplateLibrary objects as selected by the template_library config parameter.

    • If matching the spectral resolution to the galaxy data (match_resolution in config), the resolution is matched at the redshifted wavelengths of the galaxy data, adopting the input guess velocity as the redshift.

    • The template wavelength channel width is set to a fraction (1/velscale_ratio) of the galaxy data.

  • Execute the fit_method selected in config. Currently, this can only be ppxf.

  • In fit_SpatiallyBinnedSpectra():

    • Mask binned spectra, ignoring pixels masked as DONOTUSE, IGNORED, FLUXINVALID, or FORESTAR in DAP LOGCUBE file.

    • Call fit() with the data from the TemplateLibrary and SpatiallyBinnedSpectra objects.

      • If rejecting, the size of the boxcar (pixels) is set by reject_boxcar.

      • All filter_* config options are only used with fit_iter=fit_reject_filter. Do not use these options!

      • moments, degree, mdegree, and bias are passed directly to pPXF.

    • Given the template and object spectral range, determine the maximum viable fitting range for pPXF using fitting_mask().

    • Run through the specified iteration procedure, as selected by fit_iter in config; available options are set by iteration_modes().

    • Parse the pPXF results into the data table saved to the reference file.

      • Spectra without a fit are flagged as either NOFIT or FITFAILED.

      • Check if returned kinematics are near the imposed boundaries: \(v \pm 2000\) km/s from the input redshift and \({\rm d}v/100 < \sigma < 1000\) km/s, where \({\rm d}v\) is the size of the pixel (\(\sim 70\) km/s). Leads to The NEARBOUND flag in the MAPS file.

      • Flag pixels rejected by the sigma-clipping iteration.

    • Calculate the dispersion corrections:

      • First construct three spectra: (1) the optimized template; (2) the optimized template redshifted to the best-fitting velocity and with a velocity dispersion of 100 km/s; (3) the same as spectrum 2 but also convolved to the nominal object spectrum resolution.

      • Use pPXF to fit spectra 2 and 3 with spectrum 1.

      • The quadrature difference of the fitted dispersion returned for the fit to spectrum 3 and spectrum 2 is provided as the correction (STELLAR_SIGMACORR in the MAPS file)

    • Convert the pPXF velocities and velocity errors to \(cz\) velocities in km/s using convert_velocity().

  • Construct stellar-continuum BINID map. Bin IDs are the same as for the binned spectra except that any bin that does not meet the S/N limit are given a stellar-continuum bin ID of -1.