The MaStar data are obtained using the MaNGA fiber bundles feeding the BOSS Spectrographs. The raw data have the same format as the MaNGA galaxy data. The calibration scheme is also the same as MaNGA’s. Therefore, the 2d stage of the data reduction is handled by the MaNGA Data Reduction Pipeline (DRP) in nearly the same way as that for MaNGA galaxy data. This step is described in detail on this page. Basically, it processes the 2d raw data from each exposure to produce the flux calibrated, camera-combined, fiber spectra for all fibers on the plate. The only difference is in the treatment of extinction correction for standard stars. For MaNGA plates, we assume all standard stars to be behind all of the dust in their directions as they are sufficiently faint to be sufficiently far. For MaStar plates, such assumption is often invalid. Therefore, we derive the extinction for each standard star along with the flux calibration curves.
The MaStar reduction differs from the MaNGA galaxy data in the so-called “3d stage” of the MaNGA DRP. For IFU observations of galaxies, the 3d stage of the DRP produces data cubes for each fiber bundle. For MaStar observations of stars, the DRP instead produces the aperture-corrected 1D spectra for each star. This aperture correction accounts for the light lost from the central fiber due to the wavelength-dependent point spread function and the differential atmospheric refraction. Previous stellar surveys in SDSS, regardless of fiber size adopted, have similar problems when observing stars with single fibers. However, in MaStar, because we are using bundles of 2″ fibers closely packed in a hexagonal format we can use the flux ratios between the adjacent fibers to accurately constrain the PSF and position of the star relative to the fibers at each wavelength. The 3d DRP for MaStar derives this PSF-covering fraction as a function of wavelength and corrects the flux-calibrated fiber spectra to obtain the final 1-d spectra for each target star. It also combines the spectra from multiple exposures taken on the same night for each star to produce the final visit spectra, along with its uncertainty, mask, and line spread function.
The reduction pipeline is significantly updated for DR17. For details, please refer to the DR17 Data Release paper Abdurro’uf et al. (2021) and Yan et al. (in prep).