Spatial Binning

Analysis class: SpatiallyBinnedSpectra

Method definition: SpatiallyBinnedSpectraDef


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

Config files: $MANGADAP_DIR/mangadap/config/spatial_binning

Example config: vor10.ini

key                    = VOR10
galactic_reddening     = ODonnell
galactic_rv            = 3.1
method                 = voronoi
minimum_snr            = 1.0
operation              = mean
velocity_register      = False
stack_covariance_mode  = channels
stack_covariance_par   = 11
target_snr             = 10

Important class dependencies:


  • Ignore any pixels that are either masked by the boolean mask or flagged with the flags returned by do_not_use_flags().

  • Calculate the Galactic extinction (REDCORR in the DAP LOGCUBE file)

  • Using ReductionAssessment object, find spaxels with >0.8 fractional spectral coverage and above the minimum_snr in the configuration file. Only those spaxels satisfying both criteria are included in any bin.

  • Determine which spaxels to put in each bin following the method specified in the config file:

    • none (SPX binning type): No binning performed. Every selected spaxels given a unique bin ID.

    • global (ALL binning type): Bin all valid spaxels into a single spectrum.

    • radial (e.g., NRE binning type): Use the elliptical coordinates from the ReductionAssessment object to assign each spaxel to a unique radial bin. The binning annuli are defined using the center, pa, ell, radius_scale, radii, and log_step config values; see $MANGADAP_DIR/mangadap/config/spatial_binning/nre.ini for the NRE binning case. If pa, ell, or radius_scale are -1, they are replaced by pa, ell, and reff, respectively, from The DAP Datacube Configuration File.

    • voronoi (e.g., VOR10 binning type): Use the Voronoi tessellation binning algorithm (written by M. Cappellari; see vorbin) to continually accrete adjacent spaxels to reach a minimum S/N (set by target_snr in config), accounting for covariance if available, using the signal and noise measurements from the ReductionAssessment object.

  • Stack all spectra assigned to a single bin:

    • Spectra are combined following the specified operation in config. Available options are set by operation_options().

    • Account for covariance according to stack_covariance_mode and stack_covariance_par in config. Available options are set by covariance_mode_options().

    • Mask any wavelength channels in each spaxel with no unmasked pixels from the stack (maskbit set to FLUXINVALID in DAP LOGCUBE file).

  • Construct the map with the bin ID of each spaxel (BINID in MAPS file)

  • Calculate the mean signal (BIN_MFLUX in MAPS file), variance (inverse of BIN_MFLUX_IVAR in MAPS file) and S/N (BIN_SNR in MAPS file) of the stacked spectra. This is done over the same band/wavelength range as done for the individual spaxel data for the ReductionAssessment object.

  • Using the mean signal from the ReductionAssessment object, calculate the luminosity-weighted on-sky (BIN_LWSKYCOO in MAPS file) and elliptical (BIN_LWELLCOO in MAPS file) coordinates. Also calculate the unweighted coordinates; the latter are not provided in the output MAPS file.

  • Calculate the area of each bin (BIN_AREA in MAPS file), and the ratio of that area to the expected area (BIN_FAREA in MAPS file) of the binning procedure. The latter is only relevant to the radial binning, where the expected area is the area of the bin annulus.

  • Apply the Galactic reddening correction to the binned spectra, where the reddening law is defined by the galactic_reddening and galactic_rv parameters, and the E(B-V) value is taken from the DRP header keyword EBVGAL; see GalacticExtinction. The valid reddening laws are:


Internally, the DAP performs all spectral fitting on the binned spectra (termed as such even if a bin only contains a single spaxel) after they have been corrected for Galactic extinction. Therefore, the output emission-line fluxes have been corrected for Galactic extinction. However, the models and binned spectra in the output LOGCUBE file are reverted to their reddened values for direct comparison with the DRP LOGCUBE file.