Photometric redshifts for the Pan-STARRS1 survey

About this project

Tarrío & Zarattini 2020 present a method to estimate the redshift of galaxies using Pan-STARRS1 photometric data. It is an application of the algorithm proposed by Beck et al. (2016) for the SDSS Data Release 12. Our method uses a training set of 2313724 galaxies for which the spectroscopic redshift is obtained from SDSS, and magnitudes and colours are obtained from the Pan-STARRS1 Data Release 2 survey. The photometric redshift of a galaxy is then estimated by means of a local linear regression in a 5-dimensional magnitude and colour space. The method achieves an average bias of \(\overline{\Delta z_{\rm norm}}=-1.92 \times 10^{-4}\), a standard deviation of \(\sigma(\Delta z_{\rm norm})=0.0299\), and an outlier rate of \(P_o=4.30\%\) when cross-validating on the training set. Even though the relation between each of the Pan-STARRS1 colours and the spectroscopic redshifts is noisier than for SDSS colours, the results obtained by our method are very close to those yielded by SDSS data. The proposed method has the additional advantage of allowing the estimation of photometric redshifts on a larger portion of the sky \((\sim 3/4\) vs \(\sim 1/3)\).

This method was initially designed with the purpose of confirming cluster candidates of the ComPRASS catalogue (Tarrío et al. 2019). This all-sky catalogue of galaxy clusters and cluster candidates was validated by careful cross-identification with previously known clusters, especially in the SDSS and SPT footprints. Still, many candidates remain unconfirmed outside these areas. We are currently working on the confirmation of ComPRASS candidates in the PS1 area using the photometric redshifts computed with this method.

Files

Here you can download all the files that are necessary to compute the photometric redshift of a galaxy using PS1 photometry following Tarrío & Zarattini 2020. This includes:

  • ps1_training_tarrio.fits: This FITS file contains the training set, with 2313724 galaxies.
  • ps1_photoz_IDL.tar.gz: IDL scripts implementing the redshift estimation. It is composed of several files:
    • readme: describes how to use the code, and the different configuration options.
    • compute_photo_z_pan_tarrio.pro: This is the main script. It calculates the photometric redshift of a set of galaxies using the training set (ps1_training_tarrio.fits) and the photometric redshift estimation method described in Tarrío & Zarattini 2020.
    • linear_regression_zphot_tarrio.pro: This auxiliary function performs the linear regression on the 5-dimensional feature space. It is called from compute_photo_z_pan_tarrio.pro.
    • prepare_test_galaxies_tarrio.pro: This script puts the photometric information downloaded from Pan-STARRS for a given galaxy/galaxies in the appropriate format to be used as input of compute_photo_z_pan_tarrio.pro.
  • ps1_photoz_Python.tar.gz: Python scripts implementing the redshift estimation. It is composed of several files:
    • readme: describes how to use the code, and the different configuration options.
    • compute_photo_z_pan_tarrio.py: This is the main script. It calculates the photometric redshift of a set of galaxies using the training set (ps1_training_tarrio.fits) and the photometric redshift estimation method described in Tarrío & Zarattini 2020.
    • linear_regression_zphot_tarrio.py: This auxiliary function performs the linear regression on the 5-dimensional feature space. It is called from compute_photo_z_pan_tarrio.py.
    • prepare_test_galaxies_tarrio.py: This script puts the photometric information downloaded from Pan-STARRS for a given galaxy/galaxies in the appropriate format to be used as input of compute_photo_z_pan_tarrio.py.

Guidelines for using the code

The user should use compute_photo_z_pan_tarrio.pro/py to calculate the photometric redshift of a set of galaxies. The script takes as input this set, as well as the training set and different configuration parameters, and produces an output catalogue containing the photometric redshift (and error) of each galaxy in the set. The code includes several configuration options, available through different keywords. The most important ones are:

  • The user can choose the type of features to be used (aperture or Kron)
    • feat_type = ‘aper’: The 5 features to be used are the g-r, r-i, i-z, and z-y aperture colours and the r-band Kron magnitude. Note that the aperture colours are defined from the Pan-STARRS fixed aperture fluxes (see Tarrío & Zarattini 2020), and not from the Pan-STARRS aperture magnitudes.
    • feat_type = ‘kron’: The 5 features to be used are the g-r, r-i, i-z, and z-y Kron colours and the r-band Kron magnitude
  • The user can choose the subset of the full training set to be used for training when a galaxy of the input set is missing one of the features
    • train_type = ‘T5’: We use, for all the galaxies, the subset of the full training set that has the 5 features available
    • train_type = ‘T4’: For each galaxy, we take the subset of the full training set that has the same 4 features as the galaxy.
  • The standardization of the features is done by subtracting the mean and dividing by the standard deviation of the features in a given subset of the training set. The user can choose the subset to be used for calculating the mean and standard deviation to be applied in the standardization.
    • std_type ='T5': Subset of the full training set that has the 5 features available
    • std_type ='T1': For each feature, we take the subset of the full training set where that feature is available
  • The user can choose the number of neighbours that will be used in the local linear regression (k in Tarrío & Zarattini 2020) by setting the value of Nneigh.

Please refer to the README file for more details on the input/output parameters.

The script prepare_test_galaxies_tarrio.pro/py is intended for helping the user to prepare the photometric information of the galaxies in the appropriate format. For this, it takes as input the positions (ra, dec) and magnitudes (r_kron, g, r, i, z, y) of the galaxy or set of galaxies, corrects the magnitudes from reddening, computes the 5 features required by compute_photo_z_pan_tarrio.pro/py, and saves the resulting information in an appropriate FITS file. The user must indicate whether the input magnitudes are Kron or aperture magnitudes. The use of this script is not necessary, as long as the user has the photometric information in the appropriate format. For more details on the input/output parameters, please refer to the README file.

Requirements for IDL users

Requirements for Python users

Photometric redshifts

Using the code available above, we computed the photometric redshift of the PanSTARRS objects for which primary=1 and at least 4 features are available. In particular, we run the code with the following options: feat_type = 'kron', train_type = 'T5', std_type='T5', use_5feat=1, use_4feat=1, Nneigh=100.

The interested user may directly download these results here: ps1_photoz_results.tar.gz (~ 20 GB)

This compressed file contains 413 files, which correspond to the 413 files directly downloaded from the PanSTARRS1 data archive (DR2), and that cover the full survey (https://outerspace.stsci.edu/display/PANSTARRS). Each of the 413 files contains a few million objects, and for each object the following information is provided:

  • objid: Object identifier in PanSTARRS
  • ra: Right Ascension
  • dec: Declination
  • z_phot: Photometric redshift
  • z_phot_err: Estimated error of the photometric redshift
  • n_feat: Number of features that were used to compute the redshift
  • flag_extrapolate: Indicates if the redshift was calculated via an interpolation (0) or an extrapolation (1) of the training features

The naming convention of the 413 files follows that of PanSTARSS: SOV_XX.NN_zphot_final.fits, with XX=01, ..., 32 in ascending declination order. For each value of XX (i.e. each declination range) different number of files may be available, with NN starting at 00.

Acknowledging this project

If your research benefits from the use of the code or photometric redshifts of this project, please include a reference to Tarrío & Zarattini 2020 in your paper. The following acknowledgement in your paper would also be appreciated: This research has made use of the M2C Galaxy Cluster Database, constructed as part of the ERC project M2C (The Most Massive Clusters across cosmic time, ERC-Adv grant No. 340519).