4.11. Photometric Zero-Point Corrections

Photometric zero-point correction aims to correct flux measurements by band-specific photometric offsets, which can lead to systematic shifts in the observed colors, and consequently in the redshift estimates. This correction takes into account remaining zero-point calibration issues in one or more bands, and also potential mismatches between galaxy colours and the templates used to model them.

In order to derive zero-point corrections, Phosphoros follows the method implemented in the Le Phare code. It needs a training sample of the input catalog with (spectroscopic) redshift measurements. The process to derive the zero-point corrections is iterative. For each source of the training sample, Phosphoros determines the best model (with respect to the observed fluxes) fixing the redshift, and estimates in each band the offset in terms of the ratio between the predicted and observed flux, \(r=f_m^i/f^i_{obs}\). The overall offset of each band is then computed by averaging over all the sample sources. Offsets are finally used to correct photometric measurements (\(f^i_{obs}\rightarrow \langle r\rangle f^i_{obs}\)). The process is repeated until reaching a tolerance threshold for the discrepancies or a maximum number of iterations.

The averaging methods to compute the band overall offsets are the following: mean, weighted mean, median (default option) and weighted median. The weighted mean/median use the signal-to-noise value in the band as weight.

4.11.1. Photometric Corrections in the GUI

Zero-point corrections can be computed in the sub-panel 4. Photometric Zero-Point Corrections of the Compute Redshifts window.

Select the Enable Photometric Zero-point Corrections tab.

Users can compute new zero-point corrections by clicking on the Compute New Corrections button. A pop-up window will appear, as shown in Fig. 4.10.

Fig. 4.10 Computing zero-point correction with the GUI

The top of the window displays the configuration (catalog type, parameter space, filters, etc.) that will be taken into account for the zero-point corrections. Three further steps are required:

• 1. Traning Catalog Selection: select the training sample through the Browse tab, along with the column name where the spectroscopic redshifts are stored.

  The Milky Way absorption correction will be applied to the training sample following the option selected in the 1. Luminosity Filter and Extrinsic Absorption sub-panel. If the Use Galactic E(B-V) Column option is selected, the E(B-V) column must be present in the training sample; if Look-up Galactic E(B-V) in Planck Dust Map is choosen, the column PLANCK_GAL_EBV is computed and added to the training sample (or to a new file according to the user’s choice).

• 2. Algorithm: define the algorithm parameters (i.e. the Number of Iterations and the Tolerance thresold, used to decide when algorithm interations are stopped) and the Selection Method (i.e. the method to compute the overall band corrections).

• 3. Run:, choose the file to export zero-point corrections. The recommended (default) name is <Parameter Space>_<Method>. The file is in ASCII format, with the .txt extension. In the standard configuration, it is stored in the directory:

  > $PHOSPHOROS_ROOT/IntermediateProducts/<Catalog Type>/
  

Otherwise, users can select already existing files with zero-point correction values through the drop down menu below Selecting Exisiting Compatible Corrections, and visualize or modify them through the View / Edit Corrections button.

4.11.2. Photometric Corrections in the CLI

Zero-point corrections can be computed using the Phosphoros action compute_photometric_corrections (or CPC).

Action parameters can be passed with a configuration file through the --config-file action parameter. If not specified, Phosphoros reads by default the following configuration file:

/path_to_PhosphorosCore_installation_directory/conf/PhzExecutables/PhosphorosComputePhotometricCorrections.conf

Configuration files for this action can be generated through the Phosphoros GUI using the Save Config. File button present in the 4. Photometric Zero-Point Corrections sub-panel (see Fig. 4.10).

The main parameters of this action concern the training sample and the algorithm configuration parameter. An example of configuration file is following:

catalog-type=<name>
input-catalog-file=<file name>
output-phot-corr-file=<output file name>

normalization-filter=COSMOS/B_Subaru
normalization-solar-sed=solar_spectrum

source-id-column-name=<column name> # or source-id-column-index=<number>
spec-z-column-name=<column name> # or spec-z-column-index=<number>
spec-z-err-column-name=<column name> # or spec-z-err-column-index=<number>

model-grid-file=<file name>
galactic-correction-coefficient-grid-file=<file name>
dust-column-density-column-name=<column name>

phot-corr-iter-no=<value>
phot-corr-tolerance=<value>
phot-corr-selection-method=<value>

where the training sample (input-catalog-file) is searched below Catalogs/<Catalog Type> and the photometric correction output file (output-phot-corr-file) is stored in IntermediateProducts/<Catalog Type>. The recommended output name is <Parameter Space>_<Method> with the .txt extension (by default is photometric_corrections.txt). The model grid file must be the same used for the redshift computation. It is expected to be found in the directory IntermediateProducts/<Catalog Type>/ModelGrids/.

The SED normalization parameters (see SED Normalization), the column names/indices of the source ID and of the spectroscopic redshift are required by the action. On the contrary, spectroscopic redshift errors are optional and if missing, they are set to zero.

In order to include the Milky Way absorption correction, the filename of the correction coefficients grid and the column name containing the \(E_{B-V}^{MW}\) color excess must be provided, as shown above (see Galactic absorption correction in the CLI for more details).

The default values for the algorithm parameters are: 5 for the number of iterations phot-corr-iter-no; \(10^{-3}\) for the tollerance threshold phot-corr-tolerance; MEDIAN for the averaging method.

Warning

The CPC action has many more options, most of them present also in the compute_redshift action (see CLI: Computing Redshifts and its advanced functionalities). All these options should match the ones used later for the redshift estimate (by the compute_redshift action) otherwise the zero-point correction would be meaningless.