CLAMATO DR2: IGM Lyman-Alpha Forest Tomography Survey Public Data Release of Spectra and Maps

CLAMATO Data Release 2

Updated 2023 Jan 11th by Khee-Gan Lee (kglee@ipmu.jp)

Supporting paper has been published in ApJS (arXiv:2109.09660)

These are data products associated with the second data release (DR2) of the COSMOS Lyman-Alpha Mapping And Tomography Observations (CLAMATO) survey with the Keck-I telescope, which mapped 3D Lyman-alpha forest absorption at 2.05<z<2.55 within the COSMOS field. This is an updated version of the original DR2 repository (https://doi.org/10.5281/zenodo.5842842), which had accidentally left out several files. The other files that was in that repository are duplicated here.

The following is the summary of the main products:

• Source catalog (cl2020_valueadded_release_20200602.txt)
• Reduced spectra, in spec_v0 .tar.gz tarball
• Continuum-fitted 2.05<z<2.55 Lyman-alpha forest pixel data (pixel_data_v0.bin)
• Wiener-reconstructed 3D absorption map (map_2020_v0.bin)
• Reconstructed 3D matter density field using the TARDIS-II algorithm (TARDIS_CLAMATO_DR2_v0.4_smoothed.npy)
• Eigenvalues of the pseudo-deformation tensor calculated from the density field (TARDIS_eigenvalues_CLAMATO_DR2_v0.4_smoothed.npy)

 Redshift Catalog and Spectra

We provide our redshift catalog and reduced spectra obtained with Keck-I/LRIS-Blue.

The source catalog is provided in the ASCII file cl2020_valueadded_release_20200602.txt, with the following columns:

• SPECFIL: Filename of spectrum (within spec_v0.tar.gz)
• TOMO_ID: CLAMATO ID number
• GMAG: g-magnitude (AB) per Capak et al 2007 photometric catalog
• CONF: Redshift confidence grade: see https://arxiv.org/abs/1710.02894
• ZSPEC: Spectroscopic redshift as determined from CLAMATO spectrum
• QSO: QSO flag (1 if QSO, 0 if non-QSO)
• RA: R.A. in degrees (J2000)
• DEC: Dec in degrees (J2000)
• S/N_1: Estimated Lya-forest S/N at 2.05<z<2.15, -9.0 denotes no estimate
• S/N_2: Estimated Lya-forest S/N at 2.15<z<2.35, -9.0 denotes no estimate
• S/N_2: Estimated Lya-forest S/N at 2.35<z<2.55, -9.0 denotes no estimate
• S/N_RED: Estimated S/N over restframe 1250 Å < \(\lambda\) < 1350 Å, -9.0 denotes no estimate
• TOMOFLAG: Flag on whether sightline was used in tomographic map (0 for no, 1 for yes)
• EXPTIME: Exposure time on the spectrum, in seconds (aggregate)

The tarball spec_v0.tar.gz include all the reduced spectra from LRIS-Blue, with the respective filename indicated by the first column of the catalog. We decided not to make available the LRIS-Red spectra.

The individual LRIS spectra are provided in FITS format, with the following HDU Extensions:

• HDU0: Object spectral flux density, in units of \(10^{-17}\,\mathrm{erg\,s^{-1}\,cm^{-2}\,angstrom^{-1}}\)
• HDU1: Noise standard deviation
• HDU2: Pixel Wavelengths in angstroms

Pixel Data

The binary file PIXEL_DATA_v0.BIN stores the concatenated Lyman-alpha forest pixels at 2.05<z<2.55 that have been extracted from the 1D spectra and continuum-fitted. 

The first value in the binary is a 32-bit integer specifying the number of pixels (84608), followed by 5 double-precision floating point (64-bit) vectors storing the x, y, z, sigma_f, and delta_f of the pixels.

An example python script to read pixel_data is as follows:

import numpy as np with open('pixel_data_v0.bin','r') as f:    npix = np.fromfile(f, dtype=np.int32, count=1)    f.seek(4)    pixel_data = np.fromfile(f,dtype=np.float64).reshape((npix,5))

LIST_TOMO_INPUT_2020.TXT is a summary file of corresponding to PIXEL_DATA.BIN, listing the [x,y,z] position of the sightlines that contributed to the file as well as, in the final two columns, the index range that can be used to grab the relevant pixels from the concatenated pixel list.

The origin of the map coordinates is at [RA=149.89150 deg, Dec=2.0915050 deg], where x is increasing in the R.A. dimension and y is increasing in the Dec.

Tomographic Map

The Wiener-reconstructed map of the 2.05<z<2.55 IGM within the CLAMATO field is the result of applying the "dachshund" algorithm (http://github.com/caseywstark/dachshund) to PIXEL_DATA.BIN, with the configuration file INPUT.CFG . (Caveat: the version of PIXEL_DATA.BIN here is not actually the right version to directly input into the dachshund code: the first integer in this file should not be present for input to dachshund). 

The reconstructed map is MAP_2020_V0.BIN, which is a 68x56x876 = 3335808 pixel double-precision binary file. The dimension that changes fastest is the z-dimension along the line-of-sight (876 pixels per dimension), followed by the y-dimension  in increasing Declination (56 pixels per dimension) and x-dimension along the direction of increasing R.A. (68 pixels per dimension).

Each map pixel represents a 0.5Mpc/h comoving voxel of the Ly-alpha forest absorption. See the Appendix of https://arxiv.org/abs/1710.02894 for the conversion factors to assume to switch between pixel/voxel and [RA, Dec, redshift].

Note that no additional smoothing has been applied in this binary, whereas most of the visualizations in the paper have had Gaussian smoothing applied.

Reconstructed Density Field and Cosmic Web

We also release the underlying matter density field in the CLAMATO volume, estimated using the TARDIS-II algorithm (https://arxiv.org/abs/2007.15994). This is in numpy format, and can be read directly into a 3D array using Python as follows:

>>> import numpy as np >>> den=np.load('TARDIS_CLAMATO_DR2_v0.4_smoothed.npy') >>> np.shape(den) (34, 28, 438) >>> eigen = np.load('TARDIS_eigenvalues_CLAMATO_DR2_v0.4_smoothed.npy') >>> np.shape(eigen) (34, 28, 438, 3)

The 3 dimensions correspond to the R.A., Declination, and line-of-sight directions, respectively. 

Note that unlike the Wiener-filtered absorption map, the TARDIS-II reconstruction outputs are in 1Mpc/h comoving voxels. The coordinate zero-point sand line-of-sight comoving distance-redshift relationship are otherwise the same as the Wiener map. 

The eigenvalues of the pseudo-deformation tensor are in TARDIS_eigenvalues_CLAMATO_DR2_v0.4_smoothed.npy, with the same array shape as the matter density, but with an additional array dimension storing the 3 sorted eigenvalues at each point in the volume. 

For reference, we have included an iPython notebook (Tomographic_Maps.ipynb) that plots the Wiener-filtered absorption alongside the TARDIS-II densities and eigenvalues (e.g. Figure 8 of the main paper).