Spectral-Index Measurements

Analysis class: SpectralIndices

Reference root: see method_path; $MANGA_SPECTRO_ANALYSIS/$MANGADRP_VER/$MANGADAP_VER/[METHOD]/[PLATE]/[IFUDESIGN]/ref for MaNGA

Reference file: see default_paths; manga-[PLATE]-[IFUDESIGN]-[RDXQA]-[BINNING]-[CONTINUUM]-[ELFIT]-[INDICES].fits.gz for MaNGA

Optional Parameters: see Analysis Plans. The table below lists the parameters defined by SpectralIndicesDef

Key

Type

Options

Default

Description

key

str

INDXEN

Keyword used to distinguish between different spectral-index databases.

minimum_snr

int, float

0.0

Minimum S/N of spectrum to fit

pixelmask

SpectralPixelMask

Object used to mask spectral pixels

fwhm

int, float

-1

Resolution FWHM in angstroms at which to make the measurements. If -1, the measurements are taken at the native resolution of the observed spectra. If >0, the resolution should be larger than the observed resolution, and the observed data are convolved with a wavelength dependent Gaussian to match it to the provided FWHM.

compute_corrections

bool

True

Flag to compute velocity dispersion corrections to the measured indices, if possible. This requires the stellar kinematics results from the stellar continuum fitting.

absindex

str

EXTINDX

Either a string identifying the absorption-line database to use, or the direct path to the parameter file defining the database

bandhead

str

BHBASIC

Either a string identifying the bandhead/color database to use, or the direct path to the parameter file defining the database

overwrite

bool

False

If the output file already exists, redo all the calculations and overwrite it.

Important class dependencies:

  • AbsorptionIndexDB: Generalized class that provides the detailed parameters for a set of absorption-line spectral indices.

  • BandheadIndexDB: Generalized class that provides the detailed parameters for a set of bandhead (or “color”) spectral indices.

  • mangadap.proc.bandpassfilter: Provides the core functions that perform the bandpass integrals.

Algorithm:

  • Read the artifact database to setup the SpectralPixelMask object based on the artifact_mask config.

  • Setup the AbsorptionIndexDB (using absorption_indices config) and BandheadIndexDB (using bandhead_indices config) databases with the indices to measure.

  • Determine the binned spectra above the S/N limit set by the minimum_snr config.

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

  • Get the best-fitting emission-line models from the EmissionLineModel object and subtract it from them from the data; keep track of where an emission-line model is and is not defined.

  • Measure the indices using measure_indices():

    • Compute flux per frequency, needed for some indices; i.e., convert spectra from \(F_\lambda\) to \(F_\nu\).

    • Isolate which indices use each definition (\(F_\lambda\) vs. \(F_\nu\))

    • Mask any “dummy” indices.

    • For each spectrum, redshift the band definition, measure the absorption-line indices using AbsorptionLineIndices, and the bandhead indices using BandheadIndices, and save the results using save_results().

      • Part of saving the results is to determine which indices were successfully measured. Only bands that are completely masked (or empty) are flagged as NOVALUE in the output maps. I also keep track of which bands are incomplete (only partially masked).

  • Compute the velocity-dispersion corrections:

    • Get the best-fitting continuum model from the StellarContinuumModel, both with (continuum) and without (continuum_dcnvlv) the convolution with the best-fitting line-of-sight velocity distribution function (LOSVD) using fill_to_match().

    • Remeasure the indices on these two models (indx and dcnvlv_indx, respectively) and the correction based on the result using calculate_dispersion_corrections()

      • For mag unit indices, the correction is dcnvlv_indx-indx

      • For ang unit indices, the correction is dcnvlv_indx/indx

    • Any index with a bad correction is flagged as NOCORRECTION.

  • Construct spectral-index 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 spectral-index bin ID of -1.