DR9 Galaxy Target Selection

Introduction

This page provides short summaries of the target selection algorithms used for the DR9 CMASS and LOWZ samples. For the purpose of large-scale structure catalog creation, a few time-dependent details are worth noting. First, the photometric catalogs used to select targets for spectroscopy changed during the early parts of the survey. Second, the targeting algorithm run on early chunks differs.

Reproducing the target catalog

Different photometric reductions were used for different chunks (define "chunk"?) of the survey. Objects can have different photometric properties based on which imaging scan is deemed "primary" in a given reduction. These differences are driven by differences in the statistical noise in different imaging runs rather than systematic changes in the photometric pipeline. Because we are targeting objects on the steep end of the luminosity functions, this can result in substantial changes in which objects in the photometric catalogs pass the target selection cuts. For purposes of determining a completeness map of our large scale structure catalog, it is necessary to generate a parent target catalog from the imaging runs used when the targeting algorithm was run to select the spectroscopic targets. We provide this parent catalog (somewhere?). This catalog cannot be reproduced by simply applying our target selection algorithm to the final DR8 (or DR9?) imaging catalog (is this equal to main009?), and should be used as the basis of any customized LSS catalogs. We provide here a table relating the photometric reduction version used for different chunks (not sure if this is necessary? are those files even going to be public?)

Necessary for reproducing the target catalog, from comment on wiki: One detail to track -- chunk 5 and chunk 6 overlap (a bug which has now been fixed in targeting). For targets in both, we use chunk 5 photometry.

Imaging file	Photometic pipeline version	Chunks
main009:	v2_0_13	15-
main008:	v2_0_9	12-14
main007:	v2_0_5	7-11
main005:	v2_0_3	5-6
main002:	v2_0_1	3-4
comm2:	?	2
comm:	?	1

Changes to the target selection algorithm

LOWZ

There was a bug in targeting early on, where we incorrectly applied a sliding S/G cut to the LOZ sample. version 2_0_1 (chunk 3-4) and version 2_0_3 (chunk 5-6) have the wrong sliding S/G cut for the LOZ sample while version 2_0_5 (chunk 7-11) and later have it corrected. For this reason we restrict the DR9 LOZ sample to chunks greater than or equal to 7. The target selection performed by mksample exactly agrees with boss_target1 flag, except it misses one galaxy of 532677 due to a rounding error.

Insert exact cuts used; they are given in the make_samples.pro code, or on the wiki.

CMASS

For CMASS, we restrict the sample to galaxies that would pass every targeting version. It is the same as the algorithm in the final version of the targeting code, EXCEPT that the targeting code allows CMASS targets to fail the star-galaxy separation cuts IF they pass the LOWZ cuts. Since this was implemented later (we should say when?), we don't include those galaxies in the DR9 CMASS sample. There is a small disagreement remaining between boss_target1 CMASS flag and the objects in our final target flag (BR: see paragraph just above 'How to run it (on RIEMANN)'; perhaps its rounding error, or you guys can get to the bottom of it)

Insert exact cuts used; they are given in the make_samples.pro code, or on the wiki.