A Roadmap to Photometric Redshifts for LSST Data Releases

The Rubin Observatory Data Management (DM) team announces updates to the DM Tech Note “A Roadmap to Photometric Redshifts for the LSST Object Catalog” (DMTN-049).

The Rubin science community has a considerable wealth of expertise in generating photo-z catalogs and will be the primary users of the Data Release (DR) Object catalog photo-z data products. Thus, the evaluation, validation, and implementation of the DR Object catalog’s photo-z estimator(s) should be a joint venture between the DM team and the science community. This venture includes the International Programs teams making in-kind contributions related to photo-z to DM and/or the Science Collaborations.

This roadmap in DMTN-049 guides the DM team’s efforts to engage with the scientific community of data-rights holders – including the in-kind teams – in order to collectively validate and implement one or more existing photometric redshift (photo-z) estimators into the DR processing pipeline and then serve the resulting photo-z data products for the DR Object catalogs during Operations. This roadmap is a living document and will evolve over time to incorporate input from the science community.

“Letters of Recommendation” for photo-z estimators will be due 2021-09-30.
The roadmap’s first milestone is to release a formal call for “Letters of Recommendation” in early May 2021, with a deadline of Sep 30 2021. These letters will be the science community’s official channel for advocating to DM on behalf of one or more estimators or types of estimators.

“LSST Photo-z Virtual Forums” will happen monthly from February through August.
In support of this first roadmap milestone of community “Letters of Recommendation”, starting in February 2021 DM will host a series of drop-in “LSST Photo-z Virtual Forums” discussions open to all interested people (see: Invitation to Join: Photo-z Virtual Forums). These forums will focus on communicating the technical considerations for photo-z estimators, ingesting community input about the evaluation criteria to be applied to the photo-z estimators, and facilitating the development of “Letters of Recommendation”. Discussions about, e.g., the role of proposed in-kind contributions related to photo-z, schemas for inputs and outputs, training and calibration, scientific use-cases, validation and implementations will also be included.

Questions and comments about the roadmap, and/or input regarding any of its contents, are welcome as replies in this topic thread or as new topics here in the “Science - Photometric Redshifts” category on Community.lsst.org. See also the topic about the Photo-z Coordination Group which invites questions, discussion, and collaboration about any and all aspects related to LSST photo-z in this category.

LSST Photo-z Roadmap Summary

Section 1 - An introduction to photometric redshifts for the DR Object catalogs.

Section 2 - An overview of the roadmap and its timeline, including a series of “LSST Photo-z Virtual Forums” for DM to engage with the science community, and a formal call for community “Letters of Recommendation” advocating for photo-z estimators.

Section 3 - DM’s technical considerations for photo-z estimators, such as scalability, inputs and outputs, storage constraints, training, iterative development, computational processing, and implementation language.

Section 4 - Proposed preliminary evaluation criteria to be applied by DM to potential photo-z estimators, such as scientific utility, estimator type, input and output data products, scientific performance, user experience, and technical considerations. The first phase of the roadmap is to refined the evaluation criteria with input from the science community.

Section 5 - Proposed call for “Letters of Recommendation” from the science community regarding photo-z estimators (deadline 2021-09-30). From these letters, DM will apply the evaluation criteria to generate a shortlist of community-vetted photo-z estimators.

Section 6 - Proposed “Photo-z Validation Cooperative” for community participation in validating the shortlisted estimators with commissioning data (dates TBD).

The roadmap outlined in DMTN-049 also includes the proposed in-kind contributions from international partnerships, and the option of federating user-generated photo-z catalogs during Rubin operations. The roadmap’s appendices contain examples of science validation processes, aspects of the photo-z related data products (e.g., input/output schema, compression options, documentation), and a collection of scientific use-cases that motivate the evaluation criteria.