ZEBRA is a free and open-source software
distributed under the GNU General Public License 3 and can be downloaded here.

Introduction

ZEBRA combines and extends several of the classical approaches to produce accurate photometric redshifts down to faint magnitudes. In particular, ZEBRA uses the template-fitting approach to produce Maximum Likelihood and Bayesian redshift estimates based on:

• An automatic iterative technique to correct the original set of galaxy templates to best represent the SEDs of real galaxies at different redshifts;
• A training set of spectroscopic redshifts for a small fraction of the photometric sample, to improve the robustness of the photometric redshift estimates; and
• An iterative technique for Bayesian redshift estimates, which extracts the full two-dimensional redshift and template likelihood function for each galaxy.

The workflow of ZEBRA. The input files are indicated on the top. The resulting output data on the bottom. A typical ZEBRA session would take the following route: 1) Check & correction of the photometric catalog; 2) Optimization of the spectral energy distributions (templates); 3) Maximum-Likelihood fit or Bayesian estimation of the probablity distribution in template type - redshift space.

• In the photometry-check mode ZEBRA returns a catalog with calibrated photometry together with detailed information about the applied changes.
  
  
• In the template-optimization mode ZEBRA returns the corrected templates as wavelength --flux density (per unit wavelength) tables.
  
  
• In the Maximum-Likelihood mode ZEBRA returns the best fit redshift and template type together with their confidence limits estimated from constant chi^2 boundaries. ZEBRA also provides the likelihood functions for all galaxies in several output formats and the residuals between best fit template magnitude and measured magnitude for each galaxy in each filter band.
  
  
• In the Bayesian mode ZEBRA calculates the 2D-prior in redshift and template space in an iterative fashion. This prior (and, if specified, the interim prior of each iteration step) is returned. The final prior is used by ZEBRA to compute a posterior for each galaxy. The posterior can be saved as full 2D-table or in marginalized form. ZEBRA's output includes the most probable redshift and template type for each galaxy as defined by (i) the maximum of the posterior or (ii) after marginalizing over templates types or (iii) after marginalizing over redshifts, respectively.  The errors are calculated directly from the posterior.
  
  
• ZEBRA can also be employed to derive template-based k-corrections using the specified templates and filters.

a)	b)	c)
A simple Maximum-Likelihood estimate of redshifts without template optimization; (top) spectroscopic vs. photometric redshifts; (bottom) the error in the redshift estimation as function of redshift for different template types; standard deviation of dz/1+z ~ 0.043 outlier fraction ~ 2.1 %	ZEBRA's Maximum-Likelihood estimate of redshifts with template optimization. Top and bottom panel as in a). standard deviation of dz/1+z ~ 0.027 outlier fraction ~ 1.1 %	ZEBRA's Bayesian estimate of redshifts with template optimization. Top and bottom panels as in a). standard deviation of dz/1+z ~ 0.027 outlier fraction ~ 0.8 %

Please report any bugs or problems you encounter with the code to:
If you are intending to extend / improve ZEBRA it would be kind to let us know as well.

• include extinction by dust
  
  
• estimate stellar masses, ages and metallicties
  
  
• adapt the treatment of AGN templates
  
  Stay tuned!