The COSMOS Spectroscopic Redshift Compilation -- 20 years of Spectroscopy to Guide the Future

October 30, 2025

Back in March 2026, we recently posted the first data release (DR1.0) of the COSMOS Spectroscopic Redshift Compilation which encompassed 165000 redshifts for 97000 galaxies spread out over 108 individual spectroscopic follow-up programs. These measurements constitute 20 years worth of hard work and dedication from our community bundled together in a single, comprehensive data product.
 
We are happy to announce that our spectroscpic redshift compilation paper has been accepted for publication in The Astrophysical Journal Supplement and includes an updated DR1.1, which comes with many new additions. This latest version contains about 488,000 redshifts for 266,000 galaxies from over 138 individual programs. The large increase is primarily due to the inclusion of DESI DR1 & DR2 as well as other programs, such as the HST-Hyperion Survey. We have also expanded the compilation to cover a 10 deg2 area in anticipation for expected wide area surveys that will target the COSMOS in the foreseeable future, in particular, the Roman Space Telescope. Included with the compilation are SED-derived properties from CIGALE and LePHARE that are based on COSMOS2020/Classic photometry.
M_specz
 
 
The compilation provides an important and crucial legacy resource that comes at a time where astronomy is becoming increasingly data-driven. Next-generation surveys will be providing us with more spectroscopic datasets and redshift measurements such that it is becoming even more imperative that we have such resources such as the COSMOS Spec-z Compilation to keep track of data we have on-hand. As demonstrated in our compilation paper, we can then use the combined datasets with tools such as self-organizing maps to gauge what we are currently missing in our spectroscopic census of galaxy evolution facilitating further observations and enhanced targeting strategies.
 
The compilation will be periodically updated and accessible via GitHub and can be used to support many different science cases, enable new science, and provide guidance for future next-generation surveys.
 
Specz_completeness