Targeting in APOGEE-1

Spectroscopic Target Selection and Target Flags

Every 7 deg2 APOGEE field has many more objects in it than APOGEE can observe. Targeting involves the process of selecting which objects will be observed. An object may have been targeted for spectroscopy for any one of many different purposes, or for several purposes at once.

To keep track of these reasons, SDSS uses target flags. Target flags can be used in catalog queries to select only objects that were (or were not) targeted for a specific set of reasons.

The majority of APOGEE’s targeted sample is composed of red giant stars, selected with very few criteria — this is referred to as the “main sample” or “normal targets”. The other scientific sample component comprises the “special targets”, which include (among others) stars with measured parameters and abundances from other spectroscopic studies, cluster members, and targets submitted by one of APOGEE’s numerous ancillary science programs. Additionally, a sample of early-type stars is observed in order to measure terrestrial atmospheric contamination and remove it from all of the spectra.

In addition to the different target classes detailed below, you can read a much-expanded description in Zasowski et al. (2013).

APOGEE Target Bitmasks in DR13

If you are not familiar with bitmasks, please see our bitmask primer.

The targeting flags, also called bitmasks, can be used to identify objects that were targeted for some particular reason(s). The target bitmasks used by the APOGEE survey are APOGEE_TARGET1, APOGEE_TARGET2, and APOGEE_TARGET3 (note that this bit will not be populated for DR13).

We also include a simple bitmask called EXTRATARG in the summary data files and the CAS, that allows users to quickly select primary main survey targets only (no EXTRATARG bit set, i.e., EXTRATARG=0), avoiding ancillary targets, commissioning data, telluric correction targets, and duplicated observations.

Some examples of using bitmasks are provided below .

APOGEE Main Survey Targets

APOGEE’s primary sample of red giant stars is selected using H-band magnitude limits and a dereddened (J-Ks)0 color limit.

The magnitude limits are determined by the number of complete visits (hence, amount of total exposure time) expected for a group of stars. In a given field, some stars will be observed every time APOGEE visits the field, and some will only be observed for a subset of the visits. A group of stars observed together in the same visit(s) is called a cohort, which comes in one of three types (short, intermediate, long), depending on how many visits it is expected to span. The type of cohort in which a star is included in a given design is recorded in the APOGEE_TARGET1 targeting flag as “APOGEE_SHORT”, “APOGEE_INTERMEDIATE”, or “APOGEE_LONG” (bit 11, 12, or 13), and the magnitude limits are chosen such that the faintest stars in each cohort will have spectra with a final (combined) S/N of 100 per pixel.

H Magnitude Limits:
 
Number of Visits
H Magnitude Range
1
7.0 < H < 11.0
3
7.0 < H < 12.2
6
12.2 < H < 12.8
12
12.8 < H < 13.3
24
12.8 < H < 13.8

For the dereddened color selection, we apply a reddening correction to each potential target based on its E(H-4.5 μm) color excess (using the RJCE method, Majewski et al. (2011), if 4.5μm photometry is available from either Spitzer or WISE missions) or on its E(B-V) reddening value in the Schlegel et al. (1998) maps. The method used for a given star is encoded in the star’s APOGEE_TARGET1 bitmask: “APOGEE_IRAC_DERED” (bit 3), “APOGEE_WISE_DERED” (bit 4), “APOGEE_SFD_DERED” (bit 5), or APOGEE_NO_DERED (bit 6) if no reddening corrections were applied (the latter is largely confined to telluric absorption calibrators and stars on certain commissioning plates). Comparison to stellar atmospheric and Galactic stellar population models indicate that within APOGEE’s typical magnitude range, a color limit of (J-Ks)≥ 0.5 substantially reduces the dwarf contamination in the final sample. (See note on the halo fields below.)

Stars with (J-Ks)≥ 0.5 and an H mag within the relevant limits are then sampled in a way that closely approximates a random draw within each cohort. Note that the final total magnitude distribution of spectroscopic targets in a field may differ significantly from the distribution of candidates, since the former also depends on the number of each type of cohort in the field as well as on the fraction of APOGEE’s science fibers allotted to each type of cohort.

In many of APOGEE’s halo fields, we use Washington (M and T2) and DDO51 photometry to classify stars as dwarfs or giants prior to their selection as spectroscopic targets (in addition to the reddening and magnitude limits). This is done to increase the selection efficiency of giant stars in these particular fields, which have an intrinsically higher dwarf fraction in APOGEE’s magnitude range than the disk and bulge fields do. Stars targeted as photometrically classified giants are flagged as “APOGEE_WASH_GIANT”(APOGEE_TARGET1 bit 7) and are prioritized over photometrically classified dwarfs (“APOGEE_WASH_DWARF”, APOGEE_TARGET1 bit 8).

APOGEE Calibration and Cluster Targets

In addition to the primary target sample, a number of stars with published stellar parameters or chemical abundances derived from (usually optical) spectroscopy are observed in order to calibrate the ASPCAP pipeline. These targets are flagged in APOGEE_TARGET2 as “APOGEE_STANDARD_STAR” (bit 2). A concerted effort was made to target a large number of spectroscopically confirmed giants and supergiants in the Galactic bulge; these have APOGEE_TARGET2 flags “APOGEE_BULGE_GIANT” and “APOGEE_BULGE_SUPER_GIANT” (bits 11,12) set, respectively.

Some stars in the Galactic halo with data from the SEGUE surveys are also deliberately targeted in order to compare the APOGEE and SEGUE stellar analysis pipelines; these are indicated in APOGEE_TARGET1 as “APOGEE_SEGUE_OVERLAP” (bit 30). (This flag is used only for stars that used SEGUE information in their targeting. There are additional SEGUE stars that were serendipitously observed by APOGEE, particularly in the clusters, that are not flagged as SEGUE overlap targets.) Similarly, stars observed as part of the Gaia-ESO survey were also targeted, indicated with APOGEE_TARGET2 “APOGEE_GES_CALIBRATE” (bit 20) set.

A small number of stars were deliberately observed pre- and post-commissioning, to check for any systematic offsets in the derived RVs and stellar parameters; these are flagged with the APOGEE_TARGET2 “APOGEE_BULGE_RV_VERIFY” (bit 21) flag.

APOGEE targeting in stellar clusters falls into two categories. One is the targeting of well characterized clusters (mostly globular), whose members are selected as those confirmed by existing abundance, proper motion, and/or radial velocity measurements. Stars observed for this reason are flagged in APOGEE_TARGET2 as “APOGEE_CALIB_CLUSTER” (bit 10). The other category is the targeting of candidate members of poorly studied or unverified clusters, and candidate members of well studied clusters identified solely by their spatial proximity to the cluster or their position relative to the cluster locus in a color-magnitude diagram. Stars in this category are flagged as “APOGEE_SCI_CLUSTER” in APOGEE_TARGET1 (bit 9).

APOGEE Ancillary, Sgr, and Kepler Special Targets

Targets observed as part of APOGEE’s approved ancillary science programs are indicated in APOGEE_TARGET1 as “APOGEE_ANCILLARY” (bit 17) in addition to a bit set in APOGEE_TARGET1 or APOGEE_TARGET2 for each individual program. See those tables for the complete list and the APOGEE ancillary program pages (as well as Zasowski et al. 2013) for the relative prioritizations and an expanded description of each program’s goals and target selection.

Stars identified by radial velocities as members of the Sgr dwarf spheroidal galaxy were flagged as “APOGEE_SGR_DSPH” (APOGEE_TARGET1 bit 26).

APOGEE targeted several different classes of stars in the Kepler field. The majority of these are red giants that either had long-cadence data suitable for seismology studies or were part of Kepler Guest Observer programs to obtain such data, for the expanded ancillary program on ages of red giants. These were flagged as either DISK_RED_GIANT (bit 22) or APOGEE_KEPLER_SEISMO (bit 27) in APOGEE_TARGET1 . Also flagged as APOGEE_KEPLER_SEISMO are the subset of the bright dwarfs and subgiants from Chaplin et al. 2014 with short cadence data also observed by APOGEE. See Zasowski et al. (2013) and Pinsonneault et al. (2014) for a fuller description of the APOGEE+Kepler program. Please note that the ANCILLARY flag is only set for a fraction of the stars in this program.

We took advantage of the APOGEE pointings in the Kepler field to observe any bright (H<11) KOIs that were known at the time that the targets were selected. Stars that were targeted because they were Kepler planet-host candidate stars are flagged in APOGEE_TARGET1 as “APOGEE_KEPLER_HOST” (bit 28). In addition, the ancillary programs Kepler Cool Dwarfs and Eclipsing Binaries targeted stars in the Kepler field to take advantage of the exquisite lightcurves.

Note that these flags specify why a star was targeted by APOGEE, not its ultimate classification. For example, several stars that were targeted under the ancillary program Kepler Cool Dwarfs turned out to be red giants that could be seismically detected. Some stars targeted because they were likely members of the open clusters NGC 6791 or NGC 6819 also have seismic parameters. On the other hand, some stars that were targeted as seismic targets because they were included in a Kepler Guest Observer program did not receive a long enough timeseries before the failure of Kepler‘s reaction wheels to reliably determine seismic parameters.

APOGEE Telluric Correction and Sky Targets

APOGEE’s observed wavelength range contains a number of contaminating spectral features from Earth’s atmosphere, such as CO2, H2O, and CH4 absorption bands and OH airglow emission lines. The APOGEE reduction pipeline ( Nidever et al 2015 ) attempts to remove these features using observations of hot stars to characterize the telluric absorption.

The stars used for telluric absorption correction are chosen from amongst the bluest stars in the field (uncorrected for reddening), spatially distributed as evenly as possible. Generally the same set of about 35 stars is used for each visit to a given field. These targets are flagged as “APOGEE_TELLURIC”in APOGEE_TARGET2 (bit 9).

The data reduction pipeline also uses observations of star-less sky (i.e., sky positions with no 2MASS sources within 6 arcsec) to monitor the airglow and other emission. These “targets” are flagged in APOGEE_TARGET2 as “SKY” (bit 4). The resulting sky spectra are available, but only through individual exposure ap1Dvisit frames, since sky subtraction is performed on an exposure-by-exposure basis.

Sample bitmask usage

As an example, to select stars observed as part of APOGEE’s “normal” sample (regardless of whether they overlap an ancillary or other special sample), look for stars identified as “APOGEE_SHORT”, “APOGEE_INTERMEDIATE”, or “APOGEE_LONG” cohort members (see below; APOGEE_TARGET1 flags 11, 12, 13). For example, using SQL syntax:

 ((apogee_target1 & power(2,11)) != 0) | ((apogee_target1 & power(2,12)) != 0) | ((apogee_target1 & power(2,13)) != 0)

or, using predefined constants in the CAS:

 ((apogee_target1 & dbo.fApogeeTarget1('APOGEE_SHORT'))  != 0) | ((apogee_target1 & dbo.fApogeeTarget1('APOGEE_INTERMEDIATE')) != 0) | ((apogee_target1 & (dbo.fApogeeTarget1('APOGEE_LONG')) != 0)

Another example might be to select stars that were targeted because they have pre-existing stellar parameter values in the literature. These targets have “APOGEE_STANDARD_STAR” (APOGEE_TARGET2 bit 2) set, corresponding to the SQL requirement:

 (apogee_target2 & power(2,2)) != 0 or (apogee_target2 & dbo.fApogeeTarget2('APOGEE_STANDARD_STAR')) != 0

Or, one may be interested in removing all of the telluric calibrator targets from a given sample. This can be done by restricting the sample to all stars meeting the requirement that the APOGEE_TARGET2 flag “APOGEE_TELLURIC” (bit 9) is not set:

 (apogee_target2 & power(2,9)) = 0

Because objects can (and often do) have multiple targeting bits set, selections can be made on multiple criteria. For example, to identify globular cluster members that are also photometrically classified as giants, impose the requirement that APOGEE_TARGET2 flag “APOGEE_CALIB_CLUSTER” (bit 10) and APOGEE_TARGET1 flag “APOGEE_WASH_GIANT” (bit 7) both be set:

 ((apogee_target2 & power(2,10)) != 0) && ((apogee_target1 & power(2,7)) != 0)

Of course, we cannot list all of the possible queries here. The set of flags one needs to check depends on the particular query requirements. The tables at APOGEE_TARGET1 and APOGEE_TARGET2, along with the text below, summarize what the available target flags mean.

Targeting-Related Files

DR13 includes four files containing useful information on stars (both targeted and untargeted) in APOGEE’s fields and on the fields themselves, the designs, and the plates in which the targets are organized. (See further details on this organizational structure in Zasowski et al. 2013.)

apogeeObject
contains photometric and proper motion information on the candidate targets in the APOGEE fields.
apogeeField
contains information on the spatial fields, like central coordinates.
apogeeDesign
contains information on the individual designs, or groups of stars, like color limits and cohort fiber allocations.
apogeePlate
contains information on the physical plates.

Caveats and Special Notes

Overlap with MARVELS

Some of APOGEE’s earliest observations used plates shared with the MARVELS survey, with a potential small impact on the final APOGEE sample in these fields because the MARVELS targets were prioritized ahead of the APOGEE targets. Though the number of stars affected by this sharing is extremely small, a complete list of shared fields and designs is given in Table 6 of Zasowski et al. (2013).

Incorrect APOGEE IDs in the Pleiades and M54SGRC1 fields

An error resulted in incorrect APOGEE IDs for a number of stars in these fields. Lists of these IDs and the correct 2MASS identifier for the stars can be found here for the Pleiades and here for the M54SGRC1 fields. Note that because of the way the final data files are processed, one or the other of these IDs may be used. For example, the summary allStar and allVisit files use the correct 2MASS identifier, with the incorrect one preserved in the REDUCTION_ID tag, while the spectra and ASPCAP files (e.g., apStar and aspcapStar) use the incorrect ID as the primary identifier.

Special Selections and Fiber Allocations in Long Halo Fields

Due to the paucity of stars, targeting in many of the halo fields differed slightly from that in the disk and bulge fields, primarily in the use of Washington and DDO51 photometry to prioritize giants over dwarfs, as described above. Furthermore, the allocation of fibers among short, medium, and long cohorts in the long fields is set by the magnitude distribution of high-priority targets (e.g., globular cluster members). The details of these selections and fiber distributions can be found here for the following fields:

120-04 000+30 000+60 000+75 010+60 015+30
020+60 030+30 030+60 040+45 040+60 045+30
046+48 049+62 050+60 058+57 060+30 060+56
060+60 060+75 062+62 070+54 070+60 075+30
075+35 075-45 080+45 080+60 082+35 086+77
088+36 089+70 090+30 090+60 090+75 090-45
100+60 105+30 105-45 110+60 120+30 120+45
120+60 120-45 130+60 135+30 135-45 150+30
150+60 150-45 165+30 165-45 170+60 180+30
180+60 180-45 186+42 190+60 195+30 200+45
200+60 210+30 210+60 220+60 221+84 225+30
228+60 240+30 240+60 240+75 253+75 255+45
260+40 260+60 261+48 266+44 270+60 270+75
277+76 279+66 280+49 286+44 300+60 300+75
310+60 320+45 320+60 325+76 330+60 330+75
340+60 350+60 M107 M13 M3 M53
M5PAL5 M92 N1333 N4147 N5466 N5634SGR2
N6229 SGR1 VOD1 VOD2 VOD3

Proper Motion Catalogs

Many special targets (especially for the ancillary programs) have proper motions gathered from catalogs other than the ones in NOMAD (used for the main survey targets), particularly when high-accuracy proper motions were required. Here we give the references for these sources, identified by the PM_CAT tag for each star.

 
Target Proper Motion Catalogs
Catalog
Reference
KIC, KEPLER
(Kepler Input Catalog) Brown et al., 2011, AJ, 142, 112; see MAST site
CPS
(California Planet Survey) Target source only, proper motions from van Leeuwen, 2007, A&A, 474, 653
LSPM, LSPMN,
LSPM-N, LSPMd
Lépine & Shara, 2005, AJ, 129, 1483
LG11
Lépine & Gaidos, 2011, AJ, 142, 138
N188
Platais et al., 2003, AJ, 126, 2922
Schodel+09
Schödel et al., 2009, A&A, 502, 91
CRD, JNKS, LUCY,
MRTH, VSINI
Target source catalogs only; proper motions drawn from LSPM or LG11 (above)
Faherty_2009
Faherty et al., 2009, AJ, 137, 1
Jameson_2008
Jameson et al., 2008, MNRAS, 384, 1399
Kraus07
Kraus & Hillenbrand, 2007, AJ, 134, 2340
MTLS
Bonfils et al., 2005, A&A, 442, 635
PPMXL
The PPMXL Catalog

RV Verification Plate

One bulge field (006+00) was visited both before and after commissioning, with nearly identical targets, to confirm the robustness of the RV measurements. Targets on the second design are flagged in APOGEE_TARGET2 with bit 21.