NIRSpec spectra

JADES-DEEP Prism spectrum of GOODS-South #58975

Here you can find various ways to access uniformly-reduced public NIRSpec datasets:

• nirspec_public_v4.4.html:Searchable table of all public spectra reduced so far
• See the nirspec-merged-table-v4 post for information on working with summary tables of all of the reduced spectra and metadata
• For pannable images with various possible layers, including HST and JWST filters as well as spectral observations, see the Map View pages. To see individual spectra, pan over a source with the different Spectra overlays enabled.
• See the nirspec-data-products post for examples on working directly with the catalog and individual extracted spectra
• The root column below is a general rootname simply indicating spectra that were reduced together. For many programs that corresponds to individual MSA mask plans (e.g., RUBIES), but for some programs a single root name can correspond to multiple masks. For those roots that contain multiple mask plans, all spectra of a particular source from potentially multiple plans were co-added. This inconsistency of the co-addition treatment is due to the rolling nature of the processing of archival datasets as they became public and may be standardized in future reductions.

More documentation coming soon. The spectroscopy reduction process is described in de Graaff et al. (2024) and Heintz et al. (2023) and those should be cited if you use the DJA NIRSpec data products, along with any relevant citations to the separate surveys and program IDs for specific spectra. The msaexp software used to process the spectra is freely available.

Public NIRSpec datasets (v4)

These are the public datasets in the nirspec_public_v4.4.html compilation also described in nirspec-merged-table-v4.

• The v4 reductions significantly extend the wavelength range of the extracted spectra to regions that may suffer contamination of overlapping spectral orders. The sensitivity of the higher orders is strongly weighted toward the blue side of the spectrum for all gratings, so, in practice, relatively red galaxies often suffer relatively minor order contamination.
• As below, exposure time per spectrum t (in hours) computed from median(exptime) in the spectra files, which isn’t quite correct and double-counts exposure times for spectral extractions that overlap both NRS1 and NRS2 detectors
• “N total” is the total number of estimated unique sources in that complation.

Public NIRSpec datasets (v3)

These are the public datasets in the nirspec_graded_v3.html compilation:

• The root column is a general rootname simply indicating spectra that were reduced together. For many programs that corresponds to individual MSA mask plans (e.g., RUBIES), but for some programs a single root name can correspond to multiple masks. For those roots that contain multiple mask plans, all spectra of a particular source from potentially multiple plans were co-added. This inconsistency of the co-addition treatment is due to the rolling nature of the processing of archival datasets as they became public and may be standardized in future reductions.
• The root values are linked to individual tables summarizing the sources in that mask. Those tables have slightly different columns than the grade summary as they are generated before the redshifts are estimated from the spectra.
• As below, exposure time per spectrum t (in hours) computed from median(exptime) in the spectra files, which isn’t quite correct and double-counts exposure times for spectral extractions that overlap both NRS1 and NRS2 detectors
• “N total” is the total number of estimated unique sources in that complation.

Public NIRSpec datasets (v1)

These are the public datasets in the nirspec_graded.html compilation:

• Exposure time per spectrum t computed from median(effexptm) in the spectra files, which isn’t quite correct and double-counts exposure times for spectral extractions that overlap both NRS1 and NRS2 detectors

Full visualization

1000 NIRSpec prism spectra in one plot!

Spectral coverage comparison

One aspect of the unique power of the NIRSpec PRISM mode is the broad wavelength coverage from 0.7 < λ < 5.3 µm. The comparisons below show the wavelength coverage from WFC3 G141 slitless spectroscopy or from a ground-based near-infrared spectrograph such as Keck/MOSFIRE with its attendant atmospheric windows. Note that the comparison just shows the wavelength coverage, not differences in sensitivity and spectral resolution, which are significant between these observing modes.

NIRSpec WFC3 G141 Ground-based

For DJA questions, please use the Discussions or Issues functionality on the repository page.

For grizli or msaexp software questions or problems, please file an Issue on the respective GitHub repository pages.

• Cosmic Dawn Center
  Niels Bohr Institute
  University of Copenhagen
  Jagtvej 128, 2200 København N