DR9 bitmasks

This page details the meaning of the bitmasks provided with Data Release 9 of the Legacy Surveys. The bits are enumerated as a power (i.e. 7 written in a column of bits means two-to-the-power-of-7).

MASKBITS

These definitions apply to both the values in the coadd/*/*/*maskbits* files and to MASKBITS columns in the Tractor and sweeps catalogs. See also the legacypipe bitmask definitions.

Bit

Name

Description

0

NPRIMARY

touches a pixel that is outside the BRICK_PRIMARY region of a brick

1

BRIGHT

touches a pixel within half of the locus of a radius-magnitude relation. Set for Tycho sources with MAG_VT < 13 and Gaia stars with G < 13.

2

SATUR_G

touches a pixel that was saturated in at least one \(g\)-band image

3

SATUR_R

touches a pixel that was saturated in at least one \(r\)-band image

4

SATUR_Z

touches a pixel that was saturated in at least one \(z\)-band image

5

ALLMASK_G

touches a pixel that has any of the ALLMASK_G bits set

6

ALLMASK_R

touches a pixel that has any of the ALLMASK_R bits set

7

ALLMASK_Z

touches a pixel that has any of the ALLMASK_Z bits set

8

WISEM1

touches a pixel in a WISEMASK_W1 bright star mask

9

WISEM2

touches a pixel in a WISEMASK_W2 bright star mask

10

BAILOUT

touches a pixel in a blob where we "bailed out" of source fitting

11

MEDIUM

touches a pixel within the locus of a radius-magnitude relation. Set for Gaia stars with G < 16. The MEDIUM radius is twice the BRIGHT radius at the same magnitude.

12

GALAXY

touches a pixel in an SGA large galaxy

13

CLUSTER

touches a pixel in a globular cluster

14

SATUR_I

touches a pixel that was saturated in at least one \(i\)-band image (always zero prior to DR10)

15

ALLMASK_I

touches a pixel that has any of the ALLMASK_I bits set (always zero prior to DR10)

FITBITS

A bit-mask detailing any peculiarities regarding how a source was fit. See also the legacypipe bitmask definitions and the issue that prompted the development of FITBITS.

Bit

Name

Description

0

FORCED_POINTSOURCE

the source was forced to be type PSF

1

FIT_BACKGROUND

the source had its background levels independently fit

2

HIT_RADIUS_LIMIT

the source hit the radius limit during fitting (based on the limits for galaxy models in the Tractor code)

3

HIT_SERSIC_LIMIT

the source hit the Sersic index limit during fitting (e.g., see the Sersic model limits in the Tractor code)

4

FROZEN

the source was not fit (all parameters for the source were frozen at the reference catalog values)

5

BRIGHT

the source is a bright star (see also MASKBITS, above)

6

MEDIUM

the source is a medium-bright star (see also MASKBITS, above)

7

GAIA

the source is a Gaia source

8

TYCHO2

the source is a Tycho-2 star

9

LARGEGALAXY

the source is an SGA large galaxy

10

WALKER

fitting the source shifted its position by > 1 arcsec

11

RUNNER

fitting the source shifted its position by > 2.5 arcsec

12

GAIA_POINTSOURCE

the source was a Gaia source that was treated as a point source

13

ITERATIVE

the source was found during iterative detection

ALLMASK_X/ANYMASK_X

ANYMASK_X denotes a source that touches a bad pixel in any of a set of overlapping X-band images whereas ALLMASK_X denotes a source that touches a bad pixel in all of a set of overlapping X-band images. See, also, the legacypipe bitmask definitions. The ANYMASK and ALLMASK bit masks are defined as follows, mostly from the CP (NOIRLab Community Pipeline) Data Quality bits, which we map to the values in the table.

Bit

Name

Description

0

BADPIX

bad columns, hot pixels, etc.

1

SATUR

saturated

2

INTERP

interpolated

4

CR

single exposure cosmic ray

6

BLEED

bleed trail

7

TRANS

multi-exposure transient

8

EDGE

edge pixel

9

EDGE2

edge pixel

11

OUTLIER

marked as touching an outlier pixel by legacypipe itself

WISEMASK_W1/WISEMASK_W2

WISEMASK_W1 and WISEMASK_W2 have identical bit definitions, the only difference is that WISEMASK_W1 applies to masking in the W1 band and WISEMASK_W2 to masking in the W2 band.

Bit

Name

Description

0

BRIGHT

Bright star core and wings

1

SPIKE

PSF-based diffraction spike

2

GHOST

Optical ghost

3

LATENT

First latent

4

LATENT2

Second latent image

5

HALO

AllWISE-like circular halo

6

SATUR

Bright star saturation

7

SPIKE2

Geometric diffraction spike

Aaron Meisner's unWISE documentation details how these masks were derived.

CCD_CUTS

The survey CCDs and forced photometry files include a bitmask ccd_cuts that is used by the legacypipe code to discard CCDs that have low-quality observations before processing imaging for the Legacy Surveys. Any observations with non-zero vales of ccd_cuts are discarded prior to processing.

In this table, the column names are as they appear in, e.g., the survey CCDs file. The overall zero-points (zpt0) are:

  • \(g, r\) = [25.74, 25.52] for 90Prime (BASS)

  • \(z\) = 26.20 for Mosaic-3 (MzLS)

  • \(g, r, z\) = [25.08, 25.29, 24.92] for DECam (DECaLS)

Bit

Name

Description

0

ERR_LEGACYZPTS

Some error from the legacyzpts code (e.g., the CP had a WCS error, usually due to being very shallow)

1

NOT_GRZ

Not a \(g\)-, \(r\)- or \(z\)-band observation

2

NOT_THIRD_PIX

Mosaic-3 one-third-pixel interpolation problem

3

EXPTIME_LT_30

Exposure time < 30 seconds

4

CCDNMATCH_LT_20

Number of matching CCDs < 20

5

ZPT_DIFF_AVG

Zeropoint for CCD is different than the average for the exposure: abs(ccdzpt - zpt) > 0.25 for DECam, 0.1 for 90Prime, 0.1 for Mosaic-3

6

ZPT_SMALL

Zeropoint is too low: zpt < (zpt0 - P) where P is [0.5, 0.5, 1.0] mag for [DECam, 90Prime, Mosaic-3]

7

ZPT_LARGE

Zeropoint is too high: zpt > (zpt0 + Q) where Q is [0.25, 0.18, 0.8] mag for [DECam, 90Prime, Mosaic-3]

8

SKY_IS_BRIGHT

Sky is too high; ccdskycounts is more than [90, 150, 180] for DECam [\(g, r, z\)]; [200] for Mosaic-3 [\(z\)] (no cuts are made for 90Prime)

9

BADEXP_FILE

Exposure is listed in the bad_expid.txt file (90prime-bad_expid.txt, mosaic-bad_expid.txt or decam-bad_expid.txt)

10

PHRMS

Photometric calibration RMS is too large: ccdphrms > 0.2

11

RADECRMS

Astrometric calibration RMS is too large: rms (\(\sqrt(\mathtt{ccdrarms^2 + ccddecrms^2})\) > 0.4 arcsec for DECam; 0.2 arcsec for 90Prime; 0.1 arcsec for Mosaic-3

12

SEEING_BAD

Bad seeing measure: (seeing < 0 arcsec OR seeing > 3.0 arcsec), where, e.g., seeing = fwhm \(\times\) 0.262 for DECam

13

EARLY_DECAM

Defined as MJD < 56516 (note that DR8 started with MJD < 56730)

14

DEPTH_CUT

Depth cut code (this is run after the other cuts)

15

TOO_MANY_BAD_CCDS

Remove full exposures if more than 70% of the CCDs are cut

16

FLAGGED_IN_DES

Flagged as bad in DES

Some of the exact values used to create the CCD_CUTS bitmask are recorded in the legacyzpts code for 90Prime, Mosaic-3 and DECam.