The CatWISE2020 Catalog

Federico Marocco; Peter R. M. Eisenhardt; John W. Fowler; J. Davy Kirkpatrick; Aaron M. Meisner; Edward F. Schlafly; S. A. Stanford; Nelson Garcia; Dan Caselden; Michael C. Cushing; Roc M. Cutri; Jacqueline K. Faherty; Christopher R. Gelino; Anthony H. Gonzalez; Thomas H. Jarrett; Renata Koontz; Amanda Mainzer; Elijah J. Marchese; Bahram Mobasher; David J. Schlegel; Daniel Stern; Harry I. Teplitz; Edward L. Wright

doi:10.3847/1538-4365/abd805

1. Introduction

The CatWISE Preliminary Catalog (Eisenhardt et al. 2020), which was released via the NASA/IPAC Infrared Science Archive in 2019 August, consists of 900,849,014 sources over the entire sky selected from Wide-field Infrared Survey Explorer (WISE; Wright et al. 2010) and NEOWISE (Mainzer et al. 2014) survey data at 3.4 and 4.6 μm (W1 and W2) collected from 2010 to 2016. This data set includes four times as many exposures and spans over 10 times the time baseline of the AllWISE catalog (Cutri et al. 2013). CatWISE adapts AllWISE software to measure the sources in coadded images created by the unWISE team from six-month subsets of these data, each representing one coverage of the inertial sky, or epoch (Meisner et al. 2018b). The CatWISE Preliminary Catalog includes the measured motion of sources in eight epochs over the 6 yr span of the data, which are 10 times more accurate than those from AllWISE. The CatWISE Preliminary Catalog has been used to identify some of the coldest brown dwarfs known to date (Marocco et al. 2019, 2020; Meisner et al. 2020).

Nevertheless, further significant improvements are possible. The most important caveat for the CatWISE Preliminary Catalog is that the number of sources per square degree has relatively small variation over the sky. This is likely a consequence of the source detection methodology used for the Preliminary Catalog (Eisenhardt et al. 2020), which, while optimal for isolated point sources, results in significant incompleteness in high source density regions such as the Galactic Plane.

The CatWISE2020 Catalog addresses this issue by using an updated version of the unWISE catalog (Schlafly et al. 2019) as the detection list. In addition, the CatWISE2020 Catalog includes two more years of survey data from NEOWISE than does the Preliminary Catalog, increasing the number of epochs to 12 and the time span to over 8 yr. As a result, the CatWISE2020 Catalog has more than twice as many sources as the CatWISE Preliminary Catalog (five times as many in the Galactic Plane; see Section 2.2), and even better astrometric performance for faint sources. Figure 1 shows a comparison of CatWISE2020 total proper motions $\left({\rm{i}}.{\rm{e}}.,\ {\mu }_{\mathrm{tot}}=\sqrt{{\mu }_{\alpha }^{* \,2}+{\mu }_{\delta }^{2}}\right)$ ¹⁶ for a sample of 224 ultracool dwarfs within 20 pc from the Sun to values reported in the literature (Kirkpatrick et al. 2021, and references therein). The agreement is excellent across the whole motion range, and across a broad range of magnitudes since the sample includes objects as bright as W2∼7.3 mag and as faint as W2 ∼ 16.7 mag. The CatWISE2020 Catalog is therefore an excellent resource to identify ultracool dwarfs in the solar neighborhood.

**Figure 1.** A comparison of CatWISE2020 total measured motion to values reported in the literature, for 224 ultracool dwarfs within 20 pc of the Sun. All have S/N ≥ 3 CatWISE2020 motion measurements. Error bars are typically smaller than the symbols.
Download figure:
Standard image High-resolution image

Eisenhardt et al. (2020) presents a detailed description of the CatWISE Preliminary Catalog. Here we describe updates to the processing steps used for the CatWISE2020 Catalog relative to the Preliminary Catalog (Section 2), assess the astrometric and photometric performance of CatWISE2020 Catalog using comparisons to Gaia and Spitzer data (Section 3), and provide information on accessing the CatWISE2020 data products (Section 4). The Appendix summarizes known issues in the CatWISE2020 Catalog.

2. CatWISE2020 Processing Updates

A full description of the processing steps for the CatWISE Preliminary Catalog is given in Eisenhardt et al. (2020), and we describe changes for CatWISE2020 processing here. Coordinates in the CatWISE2020 Catalog are in the ICRS system at epoch MJD = 57170 (2015 May 28), while in the CatWISE Preliminary Catalog they are in J2000 at epoch MJD = 56700 (2014 February 12).

2.1. unWISE Coadds

The unWISE coadds follow the atlas tile footprint established by the WISE All-Sky Release, dividing the sky into 18,240 overlapping square tiles, each ∼1.56 deg on a side, aligned with the R.A. and decl. coordinate grid. Each tile is identified by a coadd_id in the form RRRRsDDD, where RRRR is the tile center R.A. in deci-degrees, s is the tile center decl. sign, with "p" for "+" and "m" for "–," and DDD is the tile center decl. in deci-degrees. For example, the tile containing the South Ecliptic Pole and centered near R.A. = 89.0 deg and decl. = −66.7 deg is 0890m667. For further detail, we refer the reader to section II.1.a in the Explanatory Supplement to the AllWISE Data Release Products (Cutri et al. 2013).

The CatWISE2020 data products are based on the combination of W1 and W2 exposures in the two sky coverages used for the AllWISE data release (Cutri et al. 2013) and in the 10 additional sky coverages from the NEOWISE 2019 Data Release,¹⁷ while the CatWISE Preliminary data products use AllWISE and the six additional sky coverages from the NEOWISE 2017 Data Release.¹⁸ The full-depth unWISE coadds from Meisner et al. (2018c) are used for both source detection and aperture photometry in the CatWISE Preliminary Catalog. The CatWISE2020 pipeline uses the full-depth unWISE coaddition of the AllWISE and NEOWISE 2019 Data Release for aperture photometry, while source detection is described in Section 2.2.

The CatWISE Preliminary pipeline used the 8 individual unWISE epoch coadds from Meisner et al. (2018d) for point source photometry and astrometry, with an adjustment to the world coordinate system (WCS) for the AllWISE epochs, as described in Eisenhardt et al. (2020). The CatWISE2020 pipeline uses 12 unWISE epoch coadds constructed using the methodology given in Meisner et al. (2019). We assumed that the adjustment to the AllWISE epochs had been applied to this set of 12 unWISE coadds and did not discover the adjustment was not applied until after the CatWISE2020 Catalog and Reject Table had been generated and imported into the IRSA database. This results in small astrometric offsets for sources in the CatWISE2020 Catalog, for which we provide a table of corrections (see Section 3.3).

2.2. Detection

The histogram of the number of detected sources per square degree in the CatWISE Preliminary Catalog is narrow (Figure 2). This means that the source density is no higher in the Galactic Plane than at the Galactic poles; in fact, it is slightly lower in the Galactic Plane, as the upper left panel of Figure 3 shows. The unWISE Catalog (Schlafly et al. 2019) uses a crowded-field point-source photometry code called crowdsource (Schlafly et al. 2018) which detects far more sources in high-density regions.

**Figure 3.** CatWISE Preliminary and CatWISE2020 Catalog (top) and Reject Table (bottom) source density, plotted in Galactic coordinates. See Section 2.4 for details.
Download figure:
Standard image High-resolution image

For the CatWISE2020 Catalog we therefore decided to replace the CatWISE Preliminary Catalog detection step, which uses MDET, the Multiband Detection software of Marsh & Jarrett (2012), with an updated version of the unWISE Catalog (hereafter UUC).¹⁹ The unWISE Catalog of Schlafly et al. (2019) is based on the NEOWISE 2018 Data Release, while the UUC used for the CatWISE2020 detection list is based on the NEOWISE 2019 Data Release. The unWISE Catalog measures source fluxes and static positions in the full-depth coadded image, with the measurements carried out independently in W1 and W2, while the CatWISE2020 pipeline characterizes sources jointly in both bands, measuring their fluxes, positions, and motions in epoch coadds.

CatWISE2020 processing begins with the band-merged UUC, where photometry for W1 and W2 sources within 24 is matched. Sources are measured in order of decreasing signal-to-noise ratio (S/N) within each ∼1.56 deg × 1.56 deg tile on the sky. To generate a single S/N-ordered detection list from the unWISE photometry in both bands, the CatWISE2020 pipeline determines an average flux uncertainty σ_avg for each band from the mean of the flux uncertainty of sources in the tile whose flux F is within ± 0.5% of the median flux in the band. The S/N in each band is calculated from F/σ_avg, and the W1 and W2 S/N values are root-sum-squared to determine a combined S/N for each source in the tile.

Figure 2 shows the resultant histogram of source densities for the CatWISE2020 Catalog and Reject Table compared to the CatWISE Preliminary Catalog and Reject Table, and Figure 3 compares the distribution of source density over the sky. With more than twice as many sources overall, the CatWISE2020 Catalog has a higher source density across the whole sky. The catalog source density still drops in the central part of the Galactic Plane and the Galactic Center. Sources are successfully detected in these regions, but a significant fraction of them fail one or more of the criteria for inclusion in the catalog, and therefore go in the Reject Table. Further details on the selection criteria and a more in depth discussion of the resulting source density are presented in Section 2.4.

One feature arising from the use of the UUC is the fact that some nearby, resolved galaxies are split into multiple pieces, leading to spurious localized overdensities of sources in the CatWISE2020 Catalog. A detailed discussion of the crowdsource treatment of extended galaxies is given in Schlafly et al. (2019 Sections 4.4 and 6.8) and here we briefly summarize the most important points. Splitting of extended galaxies only happens when the angular size of the galaxy exceeds several arcseconds, due to the 6" FWHM of WISE. Large galaxies listed in the HyperLEDA catalog (Makarov et al. 2014) received special treatment in the unWISE catalog. Elliptical regions around these galaxies are flagged in the unWISE bit masks (Meisner et al. 2019), and crowdsource rejects candidate new sources in these regions if they significantly overlap with a neighboring source. Galaxies in the HyperLEDA catalog are therefore less likely to be split into multiple detections, but heterogeneity in the catalog results in a somewhat sky position-dependent effectiveness of this mitigation. Ultimately, this is a trade off between completeness (specifically, the ability to successfully deblend real, blended sources) and reliability (specifically, the need to minimize the fragmenting of resolved sources). Completeness and reliability of the CatWISE2020 Catalog are discussed in further detail in Section 3.1.

2.3. Source Measurement and Uncertainties

As for the CatWISE Preliminary Catalog, the CatWISE2020 pipeline used an adapted version of the WPHOT software developed for the AllWISE pipeline to carry out source photometry, astrometry, and motion estimation, processing separately epochs taken with ascending versus descending survey scan directions, and then merging the results (Eisenhardt et al. 2020). For the CatWISE2020 pipeline, however, we did not allow WPHOT to add new sources ("active deblending") to improve the χ² of the fit to a source, as the crowdsource algorithm (Section 2.2) already provides a much more complete set of detected sources than does the MDET algorithm used for the CatWISE Preliminary Catalog.

In the CatWISE Preliminary Catalog, a minimum motion uncertainty of 10 mas yr⁻¹ was enforced, while for the CatWISE2020 Catalog, this floor was reduced to 1 mas yr⁻¹. This affects the χ² values when comparing Gaia and CatWISE2020 motions (see Section 3.3.2).

2.4. The CatWISE2020 Catalog

The final CatWISE2020 data products consist of a Catalog and a Reject Table. Like the CatWISE Preliminary Catalog, CatWISE2020 Catalog sources are required to:

(1)
be from the tile where that source is farthest from the tile edge (i.e., flagged as "primary") and
(2)
have W1 S/N ≥ 5 with no identified artifacts (a value of 0 in the left character of ab_flags) or
(3)
have W2 S/N ≥ 5 with no identified artifacts (a value of 0 in the right character of ab_flags).

There are 1,890,715,644 sources that meet these criteria. The 341,799,385 sources that fail to meet these criteria go into the Reject Table. Individual tile reject tables typically contain 11,000 sources, although near the Galactic Center they can contain over 120,000 sources due to the large number of artifacts. The larger number of artifacts in dense regions is illustrated in Figure 3 (bottom right panel), as well as in Figure 4, where we show the fraction of non-primary sources (i.e., those that fail criteria 1), and the fraction of primary sources with S/N ≥ 5 in at least one band but flagged as artifacts in the same band (i.e., those that pass criteria 1 but fail criteria 2 or 3) in three 1 deg × 15 deg strips. The three strips were chosen to trace the border of the overdensity of reject sources around the Galactic Center. They are:

(1)
− 0.5 < l < 0.5 deg and 0 < b < 15 deg;
(2)
29.5 < l < 30.5 deg and 0 < b < 15 deg;
(3)
− 0.5 < b < 0.5 deg and 35 < l < 50 deg.

**Figure 4.** Fraction of non-primary sources (black), and sources with S/N > 5 in at least one band but rejected because they are flagged as artifacts in the same band (blue and red), in three 1 × 15 deg strips.
Download figure:
Standard image High-resolution image

The fraction of sources rejected on account of being flagged as artifacts grows rapidly as a function of Galactic latitude, spiking at b < 6 deg. The incidence of artifacts is also a function of Galactic longitude, with a clear increase toward the Galactic Center. The increase in this case is not as dramatic as the increase as a function of b. The fraction of non-primary sources spikes whenever our chosen strips cross a tile boundary, which is defined to be parallel to the equatorial grid, and, therefore, is slanted when converted to Galactic coordinates. There is no net trend for the fraction of non-primary sources with either l or b, as expected since the overlap region between tiles does not vary significantly in this part of the sky.

Other noteworthy features of the source density map for the Reject Table (Figure 3, bottom row), are the two very dense circles corresponding to the celestial poles. This is purely a geometric effect, resulting from the increasing overlap between adjacent tiles as a function of decl. While tiles on the celestial equator overlap only by ∼7%, those near the poles overlap by as much as ∼56%. The number of duplicate sources grows accordingly, and since duplicates are removed from the catalog and placed in the Reject Table, the resulting source density appears artificially high.

The source density in the CatWISE2020 Catalog increases around the ecliptic poles (l ∼ 96 deg, b ∼ 30 deg, and l ∼ 276 deg, b ∼ −30 deg). This is a consequence of the greater depth of the WISE data around the ecliptic poles, which is in turn a result of the survey strategy of WISE (Wright et al. 2010).

The individual CatWISE2020 Catalog and Reject files for the 18,240 tiles were transferred to the NASA/IPAC Infrared Science Archive²⁰ (IRSA), where they were merged into the IRSA database. Information regarding access to the catalog is provided in Section 4.

2.5. CatWISE2020 Catalog Description

There are 187 formatted columns of information about each source in the CatWISE2020 Catalog. The CatWISE2020 Reject Table adds a column to indicate whether the source is primary in its tile (see Section 2.4). Descriptions of the columns can be obtained from IRSA.²¹ Most of the columns have the same names as in the AllWISE Catalog, and are described in Section II.1.a of the AllWISE Explanatory Supplement (Cutri et al. 2013). Table A1 of Eisenhardt et al. (2020) provides information about selected columns in the CatWISE catalogs that augments the information provided by IRSA.

The Galactic coordinates (glon and glat) for sources in the CatWISE Preliminary Catalog were calculated incorrectly and were not included in the IRSA release. In the CatWISE2020 Catalog, these columns are more accurate and now are included in the IRSA release but should not be used for astrometry. Two columns (w1fitr and w2fitr) remain excluded from the IRSA release of CatWISE2020 data products, for reasons explained in Table A1 of Eisenhardt et al. (2020). Finally, the CatWISE2020 Catalog and Reject Table include a new column (unwise_objid) which provides the updated unWISE Catalog (Section 2.2) identification corresponding to the source. Note that these identifications end in "r02" to avoid confusion with unrelated sources in the unWISE Catalog of Schlafly et al. (2019).

CatWISE2020 source designations should have the prefix CWISE for objects in the CatWISE2020 Catalog, and CWISER for objects in the CatWISE2020 Reject Table. The designation for each source, based on its coordinates for the J2000 equinox following the IAU truncation convention and without the leading CWISE or CWISER prefix, is given by the field source_name. For example, the quasar 3C 273 is CWISE J122906.70+020308.6.

3. Performance Characterization

The additional data, longer baseline, and different detection software lead to an improvement on some key performance parameters for the CatWISE2020 Catalog with respect to the CatWISE Preliminary Catalog. Following the performance characterization strategy adopted in Eisenhardt et al. (2020), we focus on the completeness and reliability of the CatWISE2020 Catalog at both the bright (W1, W2 < 8 mag, Section 3.1.1) and faint end (W1, W2 > 12 mag, Section 3.1.2), the photometric performance (Section 3.2), and the astrometric performance (Section 3.3). Spitzer data were used as external truth for photometric comparison, while Gaia DR2 was used for astrometric comparison.

3.1. Completeness and Reliability

3.1.1. Bright Sources

The CatWISE2020 Catalog completeness and reliability for sources with W1 or W2 < 8 mag were assessed using an updated version of the WISE Bright Star List (BSL) as a truth set. The list was generated by the WISE team for artifact flagging (see Section 4.4.g.vi in the WISE All-Sky Release Explanatory Supplement; Cutri et al. 2012), and updated for our performance assessment of the CatWISE Preliminary Catalog (see Eisenhardt et al. 2020).

The CatWISE2020 Catalog completeness was determined as the percentage of BSL sources that have astrometric matches in the CatWISE2020 Catalog as a function of BSL magnitude. Differential CatWISE2020 Catalog reliability was determined as the percentage of sources that have astrometric matches in the BSL as a function of CatWISE2020 magnitude. A relatively large matching radius of 55 (corresponding to two WISE pixels) was used, to account for the poorer centroiding accuracy expected for highly saturated sources.

Figure 5 shows the results for completeness. The CatWISE2020 Catalog appears slightly less complete for bright sources than both the CatWISE Preliminary and AllWISE Catalogs. While the CatWISE Preliminary Catalog achieves ∼99% completeness in the BSL W1 ∼ 5.5–8 mag and BSL W2 ∼ 5–8 mag ranges, the CatWISE2020 Catalog has slightly lower completeness, peaking at 98% in the 5.5 < W1 < 6.25 mag range. The CatWISE2020 Catalog completeness drops sharply for stars brighter than ∼4.5 mag, in a similar fashion to the CatWISE Preliminary Catalog, falling to ∼50% by W1 ∼ 4.3 mag and W2 ∼ 3.6 mag. Missing bright stars belong predominantly in two categories: (1) bright variable stars (e.g., AGBs and Mira-type variables) and (2) blended stars in crowded, nebulous fields. Detailed assessment in two test fields (2828p000 and 2657p288, which includes the Galactic Center) reveals that the missing bright stars are either missed at the detection stage or are detected but then dropped by the PSF-fitting software because of the poor quality of the PSF fit. There are 3742 bright stars within the footprint of tile 2828p000 (l = 33.28; b = −0.54), 65 of which are missing from the CatWISE2020 Catalog and Reject Table, and 10504 bright stars in the Galactic Center tile, 637 of which are missing. Of the 65 stars missing in the 2828p000 tile, 37 are missed at the detection stage, and the remaining 28 at the PSF-fitting stage. In the Galactic Center tile, 616 stars are missed at the detection stage, and 21 stars are missed at the PSF-fitting stage. Visual inspection reveals that stars missed at the detection stage are almost exclusively partly blended stars in fields with significant nebulosity due to interstellar dust. We refer the reader to Section 4.4 in Schlafly et al. (2019) for a more detailed discussion on the impact of nebulosity on crowdsource, and the possible incompleteness in such regions. On the other hand, stars missed at the PSF-fitting stage are predominantly very bright stars (W1 ≲ 5 mag or W2 ≲ 4 mag) and we speculate that their large saturated cores result in poor PSF fitting, leading to the sources being discarded by WPHOT.

As discussed in Eisenhardt et al. (2020), AllWISE completeness remains above 90% even for stars as bright as 0.25 mag, and should therefore be the catalog of choice for bright star science that requires a complete sample.

The CatWISE2020 Catalog achieves comparable or better reliability than the CatWISE Preliminary and AllWISE Catalogs for stars in the 2.8 < W1 < 8.5 mag and 2 < W2 < 8.5 mag range, with reliability consistently above 97%, as can be seen in Figure 6.

**Figure 6.** Differential reliability of the CatWISE2020 Catalog as a function of CatWISE2020 W1 (top) and W2 (bottom), compared to AllWISE and the CatWISE Preliminary Catalog.
Download figure:
Standard image High-resolution image

3.1.2. Faint Sources

Completeness and reliability were assessed for faint sources in an area of ∼94 deg² by comparison with the Spitzer South Pole Telescope Deep Field survey (SSDF; Ashby et al. 2013).

The analysis followed the same method described in Section 3.1.1, except that a smaller matching radius of 25 was used. The results are shown in Figures 7 and 8.

**Figure 7.** Differential completeness of the CatWISE2020 and Preliminary Catalog as a function of Spitzer 3.6 μm (top) and 4.5 μm (bottom) magnitude for sources in the SSDF.
Download figure:
Standard image High-resolution image

**Figure 8.** Differential reliability of the CatWISE2020 and Preliminary Catalog as a function of W1 (top) and W2 (bottom), for sources in the SSDF.
Download figure:
Standard image High-resolution image

The CatWISE2020 Catalog consistently achieves greater completeness than the CatWISE Preliminary Catalog with typical completeness of 99% across the 12 < [3.6] < 17.1 mag and 12 < [4.5] < 17 mag range. Figure 7 shows that the 50% completeness limit for the CatWISE2020 Catalog is [3.6] = 18.3 mag, while in [4.5] the completeness remains above 55% all the way down to [4.5] = 18.1 mag, the coverage depth of the SSDF (versus 17.8 mag and 17.4 mag for the CatWISE Preliminary Catalog).

The CatWISE2020 Catalog reliability is better than 99% for sources brighter than 14.5 mag in both W1 and W2, similar to the CatWISE Preliminary Catalog. For fainter sources, the CatWISE2020 Catalog reliability is slightly worse than the CatWISE Preliminary Catalog, in particular in the 14.5–16 mag range in both W1 and W2 (and down to 16.8 mag in W1), with a reliability ∼1% below that achieved by the CatWISE Preliminary Catalog. At fainter magnitudes, the CatWISE2020 Catalog remains a fraction of a percent less reliable than the CatWISE Preliminary Catalog in W1, while it becomes a fraction of a percent more reliable than the CatWISE Preliminary Catalog in W2.

Given that in the same brightness range the CatWISE Preliminary Catalog completeness significantly deteriorates compared to the CatWISE2020 Catalog, CatWISE2020 performance is overall superior to CatWISE Preliminary for faint stars.

3.2. Photometric Properties

The CatWISE2020 Catalog photometric depth was assessed using the SSDF and the COSMOS field.

Figure 9 compares CatWISE Preliminary Catalog (left) and CatWISE2020 Catalog (right) PSF-fitting photometry to 29 radius aperture photometry from the Spitzer S-COSMOS program (Sanders et al. 2007). These observations were obtained using long integration times (20 minutes), so the S-COSMOS data are much deeper than CatWISE. The closest CatWISE source within 275 was taken as the match to the S-COSMOS source. Because the CatWISE photometry is measured via point-source fitting, S-COSMOS sources were required to have <10% flux increase between the 19 and 29 radius apertures. In addition, because the W1 band is significantly bluer than the [3.6] band, to minimize spurious color-related effects S-COSMOS sources at [3.6] were required to have −0.1 ≤ [3.6]–[4.5] ≤ 0. These are the same criteria used in Eisenhardt et al. (2020) and Cutri et al. (2012). Figure 10 presents the same comparison to photometry from the SSDF survey.

**Figure 10.** Comparison of CatWISE photometry to Spitzer photometry for the SSDF, using the same methodology as in Figure 9. The outer contour represents a source density of 10 sources per 0.05 × 0.05 mag bin, with each additional contour showing a factor of two increase in source density.
Download figure:
Standard image High-resolution image

The comparison between the CatWISE Preliminary Catalog and CatWISE2020 Catalog and Spitzer photometry is consistent in both bands, in both fields. CatWISE2020 photometry becomes increasingly fainter than Spitzer beyond 16th mag, up to ∼0.1 mag fainter in W1 at 17.5 mag, and ∼0.06 mag fainter in W2 at 16.5 mag. This effect was also observed in CatWISE Preliminary photometry. The measured scatter reaches 0.217 mag, equivalent to a S/N of 5, at [3.6] = 17.41 mag and [4.5] = 16.50 mag. Adjusting for the mean offsets in W1–[3.6] and W2–[4.5] at these magnitudes, and compensating for the different depth of the three different surveys, the S/N = 5 limits for the CatWISE2020 Catalog are W1 = 17.43 mag and W2 = 16.47 mag (versus W1 = 17.41 mag and W2 = 16.33 mag for the CatWISE Preliminary Catalog).

3.3. Astrometric Properties

3.3.1. Full-sky Astrometric Assessments

The astrometric performance of the CatWISE2020 Catalog over the entire sky was assessed by comparing to Gaia DR2 (Gaia Collaboration et al. 2018; Lindegren et al. 2018). Following the method described in Eisenhardt et al. (2020), in each tile we selected the 10 brightest sources in each 0.5 mag bin over the 10 < W1 < 17.5 mag range. This resulted in a sample of 150 sources per tile, uniformly distributed on the sky. The full sample for astrometric comparison consists of 2,735,892 sources. These sources were cross-matched with Gaia DR2 using a 55 radius (corresponding to two WISE pixels), requiring the Gaia counterpart to have measured proper motions. This returned 2,179,410 unique matches. The Gaia counterparts were propagated to the CatWISE2020 epoch (MJD = 57170) using Gaia astrometry, and then the median difference and standard deviation between the CatWISE2020 motion fit and Gaia position and motion values were computed.

As mentioned in Section 2.1, CatWISE2020 processing did not apply the WCS adjustments to the unWISE coadds for the AllWISE epochs. This results in small systematic offsets between the CatWISE2020 Catalog and Gaia DR2 in both position and motions. The R.A. and decl. offset with respect to the Gaia DR2 counterpart (i.e., (R.A._Gaia–R.A._CatWISE2020) × cos decl._CatWISE2020 and decl._Gaia – decl._CatWISE2020), as well as the proper motion difference, was computed for all 2,179,410 sources in the comparison sample. Figure 11 shows the offset distribution, revealing a roughly Gaussian distributions in R.A. and decl. (top panel), with a 1σ dispersion of 46 mas and 32 mas respectively. The proper motion offsets (bottom panel) show a more complex distribution since they are the result of the projected solar motion through the Galaxy.

**Figure 11.** The distribution of median offsets between CatWISE2020 positions (top) and proper motions (bottom) and Gaia DR2 in the 18,240 tiles.
Download figure:
Standard image High-resolution image

Figure 12 maps the systematic offsets between the CatWISE2020 Catalog and Gaia DR2, in Galactic coordinates. The systematic offset is not dependent on source density, i.e., there is no obvious degradation in high-density regions. In the very center of the Milky Way the maps appear noisier, but that is a spurious result of the low number of matches between the CatWISE2020 Catalog and Gaia DR2 rather than a real astrometric effect. All maps show patterns that are consistent with the projection of the solar motion through the Galaxy. The offsets are typically in the ±150 mas range in position, and in the ±40 mas yr⁻¹ range in proper motion. The median position and proper motion offsets in each tile are provided in a machine-readable table, and a sample is shown in Table 1. For any application that requires the study of the kinematics of a large sample of objects, we recommend using these median values to correct motions and positions on a tile-by-tile basis. The tile in which a source is measured is encoded in the first 8 characters of the source_id, given in the first column of the CatWISE2020 Catalog and Reject Table. For example, source 0000p000_b0-013835 was measured in tile 0000p000.

Table 1. Systematic Offsets between the CatWISE2020 Catalog and Gaia DR2

	Gaia DR2—CatWISE2020 Catalog
Tile	Δα^*	Δα	Δδ	${\rm{\Delta }}{\mu }_{\alpha }^{* }$	Δμ_δ
	deg	deg	deg	arcsec yr⁻¹	arcsec yr⁻¹
0000m016	7.038531 × 10⁻⁷	7.041276 × 10⁻⁷	2.211593 × 10⁻⁶	−0.01956579	0.018327646
0000m031	9.905521 × 10⁻⁶	9.920038 × 10⁻⁶	6.129016 × 10⁻⁶	−0.017017405	0.0168396
0000m046	1.5195505 × 10⁻⁵	1.5244609 × 10⁻⁵	−1.5518368 × 10⁻⁵	−0.013825462	0.016300509
0000m061	2.3382128 × 10⁻⁵	2.3515273 × 10⁻⁵	−8.676556 × 10⁻⁶	−0.014948687	0.012481616
0000m076	4.49237 × 10⁻⁵	4.532183 × 10⁻⁵	−1.4328044 × 10⁻⁵	−0.015264964	0.016823826

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as: Data Typeset image

We applied the systematic offsets presented in Table 1 to the 2,179,410 sources in the comparison sample, and computed standard deviation between the CatWISE2020 motion fit and Gaia position and motion values. The results are summarized in Figure 13.

The positional accuracy floor for bright sources approaches ∼35 mas, a factor ∼1.4 better than the performance achieved by the CatWISE Preliminary Catalog (Figure 13, top left panel). At fainter magnitudes, the CatWISE2020 Catalog also shows improved astrometric accuracy. The dispersion in the 15.0–15.5 mag bin is ∼165 mas in the CatWISE2020 Catalog, while it is ∼220 mas in the CatWISE Preliminary Catalog. In the faintest magnitude bin the dispersion in the CatWISE2020 Catalog is ∼700 mas.

The motion accuracy floor for bright stars is just over 6 to 7 mas yr⁻¹, as illustrated in the top right panel of Figure 13. At the faint end, the CatWISE2020 Catalog also shows improved performance with respect to the CatWISE Preliminary Catalog. In the 15.0–15.5 mag bin the 1σ dispersion is ∼20 mas yr⁻¹, compared to ∼26 mas yr⁻¹ in the CatWISE Preliminary Catalog. In the faintest magnitude bin the dispersion in the CatWISE2020 Catalog is ∼100 mas yr⁻¹.

The χ² panels show different performance for positions and motions. The bottom left panel of Figure 13 shows improved χ² values for bright stars in the CatWISE2020 catalog with respect to the CatWISE Preliminary Catalog, with the χ² values rising in fainter sources to the same level observed in the CatWISE Preliminary catalog. The motion χ² (bottom right panel) for bright sources on the other hand are worse in the CatWISE2020 Catalog than in the CatWISE Preliminary Catalog, gradually improving toward fainter magnitudes and reaching the same level as the CatWISE Preliminary Catalog at W1, W2∼15 mag. The position χ² are significantly improved, most likely as a result of the better deblending, while motion χ² are worse because of the smaller floor on the motion uncertainty imposed by the CatWISE2020 pipeline (Section 2.3).

To assess the possible role of increasing Gaia errors, Figure 14 shows the astrometric performance of the CatWISE2020 Catalog as a function of Gaia G magnitude (top row) and distance (bottom row). The top row plots show a behavior similar to the one illustrated in Figure 13, with accuracy floors of 30 mas and 6 mas yr⁻¹ in positions and proper motions, respectively, and accuracies of 600 mas and 60 mas yr⁻¹ at the faint end. The bottom row plots of Figure 14 on the other hand show a less intuitive trend. The very nearest distance bins are dominated by very bright, highly saturated sources. As a result, the measured performances with respect to Gaia are ∼200 mas in position and 40 mas yr⁻¹ in proper motion. The astrometric performance improves very quickly as a function of distance, with minimum dispersion in the 60–80 pc bin. Beyond 80 pc the astrometry deteriorates again, as one might expect since in more distant bins the stars appear, on average, fainter. At ∼2.1 kpc, however, the trend briefly reverses, reaching a local minimum at ∼2.6 kpc. Beyond ∼2.6 kpc the astrometric performance resumes to very gradually deteriorate as a function of distance for positions, while for motions it slightly improves with distance. This pattern is a result of the change of the brightness distribution of stars in the CatWISE2020 Catalog as a function of distance. The top panel of Figure 15 illustrates this point. The nearest bin is dominated by bright sources and as such the astrometric performance is comparable to the astrometric performance in the brightest bins in Figure 13. The brightness distribution of more distant bins peaks at fainter magnitudes and as such the astrometric performance degrades, once again following the well-demonstrated trend with brightness. Somewhat counterintuitively, however, beyond ∼2.1 kpc the brightness distribution becomes bimodal, with a growing peak at bright magnitudes. This bright peak is most likely due to red giant stars. This can be seen as well in the bottom panel of Figure 15, where the distance distribution of 10 < W1 ≤ 10.5 mag stars shows a secondary peak at d ∼ 2.6 kpc. RGB stars have −5 ≲ M_W1 ≲ −2 mag (Conroy et al. 2018), roughly corresponding to the observed W1 = 10 mag at this distance. Moving beyond ∼2.6 kpc, the top panel of Figure 15 shows that the bright peak moves toward fainter magnitudes and, therefore, the accuracy of position measurements resumes degrading. While sources in a given distance bin become on average fainter with distance, their proper motion becomes on average smaller, counteracting the brightness-dependent astrometric degradation and resulting in a slow improvement of the proper motion performance as a function of distance.

**Figure 15.** Top: the W1 magnitude distribution for CatWISE2020 Catalog sources with Gaia counterparts in four distance bins. The three nearest distance bins shown are chosen to be the three inflection points in the bottom row plots of Figure 14. The fourth distance bin is chosen to represent the brightness distribution for sources with d ≳ 2.6 kpc. Bottom: the distance distribution of sources in four magnitude bins. The distribution in the two brightest magnitude intervals show clear secondary peaks at d > 2 kpc, which are due to bright, distant RGB stars.
Download figure:
Standard image High-resolution image

The overall astrometric performance is, as one might expect, not uniform over the sky. Figures 16–19 show the 1σ dispersion in each tile with respect to Gaia positions and motion components for the full magnitude range considered (Figures 16 and 18), and in three smaller magnitude intervals (Figures 17 and 19). The maps for the full magnitude range are smooth overall, indicating a fairly constant astrometric performance for the CatWISE2020 Catalog over the majority of the sky. The main features can be easily identified—the Galactic Plane (and in particular the bulge), and the Small and Large Magellanic Clouds (SMC and LMC). In those denser regions, the astrometric accuracy for the bright stars deteriorates to ∼500 mas for positions and ∼30 mas yr⁻¹ for motions, and to ∼1000 mas and ∼200 mas yr⁻¹ (or worse) for the faint stars.

**Figure 16.** 1σ dispersion of the CatWISE2020 Catalog α (left) and δ (right) with respect to Gaia DR2, for sources in the 10 < W1 < 17.5 mag range.
Download figure:
Standard image High-resolution image

**Figure 17.** Same as Figure 16, but for three W1 magnitude ranges. Gray tiles are those where there were no sources in the CatWISE2020 Catalog in the given magnitude bin.
Download figure:
Standard image High-resolution image

**Figure 18.** Same as Figure 16, but for the proper motion components.
Download figure:
Standard image High-resolution image

**Figure 19.** Same as Figure 18, but for three W1 magnitude ranges. Gray tiles are those where there were no sources in the CatWISE2020 Catalog in the given magnitude bin.
Download figure:
Standard image High-resolution image

The motion accuracy maps of Figures 18 and 19 show additional features, already noted in the CatWISE Preliminary Catalog, which appear to be related to the WISE survey strategy and to the transition between the cryogenic and post-cryogenic phases of the mission.

The maps for the faintest magnitude interval (bottom row of Figures 17 and 19) are noisier, partly because of the low number of sources with a Gaia counterpart in each tile, and partly because the CatWISE2020 pipeline detects only brighter sources in high-density regions because of confusion noise. The astrometric performance of the CatWISE2020 Catalog remains fairly homogeneous over the sky even for the faintest sources, but the performance starts degrading at higher b as one considers fainter and fainter sources. This is most likely due to the fact that the astrometry for faint sources is more susceptible to degradation due to blending.

To further assess the astrometric performance of the CatWISE2020 pipeline on faint, red sources that are not seen by Gaia, we compared the CatWISE2020 positions and motions to those measured in the Extended Gaia-PS1-SDSS Proper Motion Catalog (GPS1+; Tian et al. 2020). The GPS1+ catalog provides positions and proper motions for ∼400 million sources with 19.0 ≲ r ≲ 22.5 mag over 3/4 of the sky, measured using the combination of Gaia, PS1, SDSS, and 2MASS data.

We cross-matched our astrometric comparison sample of 2,735,892 stars with GPS1+ with a matching radius of 55, the same radius used in the comparison against Gaia DR2. We then computed the 1σ dispersion between the CatWISE2020 Catalog and GPS1+ positions and proper motions following the same procedure adopted for our comparison against Gaia DR2. Since GPS1+ omits sources with r ≲ 19 mag, the correct matching source is often missing for CatWISE2020 sources of W1 < 14.5 mag. Therefore, for the rest of the analysis we only consider sources with W1 ≥ 14.5 mag.

Figure 20 shows the results of the comparison. The 1σ dispersion, presented in the top row, increases at fainter magnitudes, from ∼200 mas for positions and ∼20 mas yr⁻¹ for proper motions for the brightest sources, to ∼650 mas for positions and ∼90 mas yr⁻¹ for proper motions for the faintest sources. The performance at the faint end agrees well with the full-sky performance derived in the comparison to Gaia DR2. At the bright end the omitted GPS1+ sources and resulting mismatches degrade the performance relative to the Gaia DR2 comparison shown in Figure 13. The bottom row of Figure 20 shows that the median χ² increases as a function of brightness, with a clearer trend for positions. Once again, this can be partly explained by the omission of bright GPS1+ sources and resulting mismatches in the brightest bins.

3.3.2. Astrometric Assessments in Selected Tiles

We assessed the astrometric performance of the CatWISE2020 Catalog in four representative tiles. These are the same four tiles we chose for the astrometric assessment of the CatWISE Preliminary Catalog (Eisenhardt et al. 2020), and are the following:

1.
the COSMOS tile (tile 1497p015), representative of most of the sky, i.e., a field with average WISE coverage and average source density;
2.
the North Ecliptic Pole tile (NEP, tile 2709p666), a field with maximal WISE coverage and average source density;
3.
the South Ecliptic Pole tile (SEP, tile 0890m667) a field with maximal WISE coverage and high source density (the SEP tile contains part of the LMC); and
4.
the Galactic Center tile (GC, tile 2657m288) a field with average WISE coverage and maximal source density.

The astrometric performance of the CatWISE2020 Catalog was assessed following the same method described in Section 3.3.1. The results are shown in Figures 21–24. We quantitatively characterized the performance of the CatWISE2020 Catalog using the same 10 metrics used in Eisenhardt et al. (2020):

1.
σ_min and σ_μ,min are the accuracy floor for positions and motions, respectively, determined as the median dispersion with respect to Gaia DR2 in the 8 < W1, W2 < 10 mag interval. In the GC we restrict to 8 < W1, W2 < 9 mag since the astrometric accuracy starts deteriorating significantly beyond W1, W2∼9 mag (see Figure 24).
2.
W1_min, W2_min, W1_μ,min, and W2_μ,min are the W1 and W2 mag at which σ_min and σ_μ,min are exceeded by no more than 20 mas and 5 mas yr⁻¹, respectively.
3.
W1₅₀₀ and W2₅₀₀ are the W1 and W2 mag at which the accuracy on positions reaches 500 mas.
4.
W1_μ,100 and W2_μ,100 are the W1 and W2 mag at which the accuracy on motion reaches 100 mas yr⁻¹.

**Figure 21.** CatWISE Preliminary Catalog (blue) and CatWISE2020 Catalog (red) astrometric performance with respect to Gaia DR2 in the COSMOS tile (1497p015). The top row shows the 1σ dispersion between CatWISE and Gaia R.A. (specifically, ${\rm{\Delta }}\alpha {\rm{\cos }}(\delta )$ ) and decl. (left) and proper motion (right). The bottom row shows the median χ² computed taking into account CatWISE catalog uncertainties, Gaia catalog uncertainties, and the uncertainty introduced by the translation of Gaia's positions to the CatWISE epoch.
Download figure:
Standard image High-resolution image

**Figure 21.** CatWISE Preliminary Catalog (blue) and CatWISE2020 Catalog (red) astrometric performance with respect to Gaia DR2 in the COSMOS tile (1497p015). The top row shows the 1σ dispersion between CatWISE and Gaia R.A. (specifically, ${\rm{\Delta }}\alpha {\rm{\cos }}(\delta )$ ) and decl. (left) and proper motion (right). The bottom row shows the median χ² computed taking into account CatWISE catalog uncertainties, Gaia catalog uncertainties, and the uncertainty introduced by the translation of Gaia's positions to the CatWISE epoch.
Download figure:
Standard image High-resolution image

The results for the four representative tiles are summarized and compared to the corresponding CatWISE Preliminary Catalog values in Table 2.

Table 2. CatWISE Astrometric Performance Evaluation Fields

	0890m667		1497p015		2657m288		2709p666
	SEP, LMC		COSMOS		GC		NEP
l (deg)	276.5		237.3		359.8		96.4
b (deg)	−30.2		41.4		0.6		29.5
β (deg)	−89.6		−10.2		−5.4		89.6

	Prelim.	2020	Prelim.	2020	Prelim.	2020	Prelim.	2020

Exp.	7154	9230	90	140	86	133	7839	10000
#	71462	209518	58961	75639	63368	231239	61702	142975

σ_min (mas)	52.9	40.5	27.3	42.8	526.4	391.2	37.7	25.8
W1_min (mag)	11.0	11.5	12.5	14.5	8.0	8.5	12.0	12.0
W1₅₀₀ (mag)	15.1	16.0	17.0	17.5	8.4	9.0	18.5	18.0
W2_min (mag)	11.0	12.0	12.5	14.5	8.0	8.5	12.0	12.5
W2₅₀₀ (mag)	15.0	16.0	16.8	17.5	8.3	9.0	19.0	20.5

σ_μ,min (mas yr⁻¹)	7.4	8.6	8.5	8.0	22.2	20.0	7.3	7.0
W1_μ,min (mag)	14.5	13.5	13.5	14.5	9.0	9.0	15.5	14.5
W1_μ,100 (mag)	18.2	18.5	16.8	17.5	>11.0	11.5	>19.0	21.0
W2_μ,min (mag)	14.5	13.5	13.5	14.5	9.0	9.0	15.5	14.0
W2_μ,100 (mag)	>20.5	19.0	16.7	17.5	>11.5	11.5	>20.0	20.5

Note. l, b, and β are the Galactic longitude, Galactic latitude, and ecliptic latitude for the center of the tile, in degrees. Exp. indicates the number of exposures for the tile, # the number of sources (combining catalog and reject entries). The subsequent metrics are described in detail in Section 3.3.

Download table as: ASCII Typeset image

In the COSMOS tile (Figure 21), the CatWISE2020 Catalog shows improved astrometric performance with respect to the CatWISE Preliminary Catalog in all metrics except the position accuracy floor. Table 2 shows that σ_min deteriorated from 27.3 to 42.8 mas. Inspection of the top left panel of Figure 21 suggests that the asymptotic performance is actually comparable for the CatWISE2020 and Preliminary Catalog, with an accuracy floor of ∼40 mas yr⁻¹. The CatWISE Preliminary Catalog, however, shows a dip in the measured motion sigma at W1, W2∼9 mag, which drives the σ_min to low values. Because the number of sources in this brightness regime is much lower compared to the fainter brightness regime (32 objects with W1, W2 ≤ 9.5 mag), we suspect that that our metric may be affected by small number statistic fluctuations. The motion accuracy floor is not affected by the already mentioned systematic offsets, since in an individual tile the systematic offset is roughly constant over the entire tile area, and therefore does not impact the measured dispersion with respect to Gaia.

In the Galactic Center (Figure 24), the major gain of the CatWISE2020 Catalog with respect to the CatWISE Preliminary Catalog is the increased depth to which sources are recovered, with the CatWISE2020 Catalog now reaching down to 16th magnitude in W1 and W2. Most metrics also show improved or consistent performance, with the position accuracy floor (σ_min) having improved significantly as a result of the more effective deblending. W1_min and W2_min, as well as W1₅₀₀ and W2₅₀₀, are 0.5–0.7 mag deeper in the CatWISE2020 Catalog. The motion metrics show comparable or slightly improved performance. The position χ² are significantly improved as a result of the better deblending, while motion χ² are worse because of the smaller floor on the motion uncertainty (Section 2.3).

The NEP and SEP (Figures 22 and 23) show improved performance for positions, but somewhat worse performance for motions. All position metrics either improved or remained constant in the CatWISE2020 Catalog, with the exception of W1₅₀₀ in the NEP, which is 0.5 mag shallower. This metric, however, is close to the S/N 5 depth, and small number statistics noise is to be expected, because of the small number of CatWISE2020 sources that have a counterpart in Gaia DR2 at such brightness. The motion metrics, on the other hand, are somewhat degraded. While the accuracy floor is at the same level achieved by the CatWISE Preliminary Catalog, the top right panel of Figures 22 and 23 shows that the CatWISE2020 motion dispersion is higher than in the CatWISE Preliminary Catalog across the entire magnitude range considered. Position χ² are better at both poles, while motion χ² are worse, once again because of the smaller uncertainty floor.

**Figure 22.** Same as Figure 21 but for the North Ecliptic Pole tile (2709p666).
Download figure:
Standard image High-resolution image

**Figure 23.** Same as Figure 21 but for the South Ecliptic Pole tile (0890m667).
Download figure:
Standard image High-resolution image

Overall, for faint sources all metrics in the CatWISE2020 Catalog retain the clear dependence on source density and coverage that was already noted by Eisenhardt et al. (2020) in the CatWISE Preliminary Catalog. The NEP (high coverage, average density) shows better W1₅₀₀, W2₅₀₀, W1_μ,100, and W2_μ,100 than the SEP (high coverage, high density) and COSMOS (average coverage, average density), which are in turn better than the GC (average coverage, high density).

3.3.3. Fast Movers

In Eisenhardt et al. (2020) we discussed the performance of the CatWISE Preliminary Catalog on a sample of cold, fast moving brown dwarfs in the solar neighborhood. Re-assessment of the CatWISE2020 Catalog performance on the same sample of brown dwarfs reveals that the motion accuracy is comparable in the two versions of the catalog. Moreover, thanks to a more effective deblending of partly blended sources, the CatWISE2020 pipeline successfully measures WISE J163940.83–684738.6 and WISEPC J205628.90+145953.3, the only two objects in our sample of 19 test objects that were missing from the CatWISE Preliminary Catalog.

One feature common to all of the fast moving brown dwarfs considered in this analysis is that their fast motion leads to multiple (spurious) detections, since they are essentially "smeared" in the full-depth unWISE coadds used for source detection. These multiple apparitions (up to seven for the fastest objects) are then passed through our photometry- and motion-measuring software. The first detection processed by WPHOT gets accurate motion and magnitude measurements, while the subsequent detections get progressively worse measurements. Those with the lowest S/N are typically discarded in the Reject Table, but the higher S/N ones remain in the catalog. For example, the fastest moving brown dwarf, WISE J085510.83–071442.5, appears five times in the CatWISE2020 Catalog, with three of those five apparitions having motion measurements consistent with the literature values.

As briefly discussed in Section 1, the performance of the CatWISE2020 Catalog on fast moving sources was further assessed by comparing the CatWISE2020 offset-corrected total measured motion for a sample of ultracool dwarfs within 20 pc from the Sun to values reported in the literature (Kirkpatrick et al. 2021, and references therein). The full sample consists of 224 objects, all of them with a counterpart in the CatWISE2020 Catalog, and with good quality proper motion measurements (i.e., μ_tot/σ μ_tot ≥ 3). Figure 1 shows a comparison of the CatWISE2020 measurements for these 224 objects to values reported in the literature. The agreement is excellent and does not show strong dependence on total proper motion or brightness since the sample includes objects as bright as W2∼7.3 mag and as faint as W2∼16.7 mag. The 1σ dispersion is ∼29 mas yr⁻¹, while the median offset is ∼12 mas yr⁻¹. Many of the 224 objects in this sample are too faint at optical wavelengths to be seen by Gaia. Therefore, the CatWISE2020 Catalog crucially complements the ESA mission for late-type stars and brown dwarfs.

4. Data Access

The merged files for the 18,240 tiles for the CatWISE2020 Catalog and Reject Table are available from IRSA²² in the WISE/NEOWISE Enhanced and Contributed Products area. IRSA's catalog search tools allow for complex search queries. IRSA also hosts the AllWISE Explanatory Supplement (Cutri et al. 2013), which provides full details on the AllWISE processing algorithms, and includes descriptions of the AllWISE Catalog columns, many of which are applicable to the CatWISE data products. Section 2.5 provides additional information about CatWISE2020 columns.

The individual tile files have also been transferred to a data repository at the National Energy Research Scientific Computing Center²³ (NERSC) and are available in 18,240 pairs of gzipped ASCII files (one catalog and one reject file per tile) in IPAC table format, organized into 359 directories, one for each decimal degree of R.A. from 0° to 358° (there are no tiles whose ID begins with 359). Text files providing the format and a brief description of the columns in the catalog and reject files are also provided there. As for the CatWISE Preliminary Catalog, the catalog and reject files for tiles near the ecliptic poles (listed in Table 1 in Eisenhardt et al. 2020), where a single PSF per band was used for processing, include the string "opt0" in their names. Files for tiles where different PSFs were used for ascending and descending scans include the string "opt1" in their names.

Current information about CatWISE data products and links to the data on IRSA and NERSC are provided at the CatWISE website https://catwise.github.io.

We thank the anonymous referee for their careful review and thoughtful suggestions. CatWISE uses data products from WISE, which is a joint project of the University of California, Los Angeles, and the Jet Propulsion Laboratory (JPL)/California Institute of Technology (Caltech), funded by the National Aeronautics and Space Administration (NASA), and from NEOWISE, which is a joint project of JPL/Caltech and the University of Arizona funded by NASA. Characterization of CatWISE performance uses data from Gaia and from Spitzer. CatWISE is led by JPL/Caltech, with funding from NASA's Astrophysics Data Analysis Program (ADAP), and is also supported in part by ADAP grant NNH17AE75I at Lawrence Berkeley Laboratory. F.M. acknowledges support from the NASA Postdoctoral Program at the Jet Propulsion Laboratory, administered by Universities Space Research Association under a contract with NASA. F.M. also acknowledges support from grant 80NSSC20K0452 under the NASA Astrophysics Data Analysis Program. F.M. would like to thank Dongwei Fan for help with accessing the GPS1+ data via the China-VO.

Appendix: Caveats

The CatWISE2020 Catalog contains a number of features that users should be aware of. Among these are:

1.
The CatWISE2020 Catalog represents a dramatic improvement over the CatWISE Preliminary Catalog in terms of depth and completeness in the Galactic Plane. However, the CatWISE2020 Catalog astrometric performance degrades as source density increases. Figures 16–19 illustrate this.
2.
The completeness for bright sources is low (Figure 5). Users interested in complete samples brighter than 4.5 mag in WISE bands should use AllWISE. However, the CatWISE2020 Catalog has much better reliability for bright sources than does the CatWISE Preliminary Catalog (Figure 6).
3.
Small systematic offsets are present in CatWISE2020 astrometry with respect to Gaia DR2 (see Sections 2.1, 3.3, and Figure 12). Because adjustments were made to the WCS for the AllWISE epochs for the CatWISE Preliminary Catalog but not for the CatWISE2020 Catalog, these systematic offsets are different between the two catalogs. Table 1 provides the corresponding systematic corrections to the CatWISE2020 Catalog position and motion for each tile.
4.
Because of coordinate singularities, CatWISE tile point spread functions (see Section 3.2 of Eisenhardt et al. 2020) within a few degrees of the equatorial poles used an unnecessarily large range of rotation angles, resulting in smearing of these PSFs.
5.
Magnitude and position uncertainties occasionally round to 0.
6.
The aperture magnitudes are the result of averaging fluxes in the CatWISE2020 Catalog, while in the CatWISE Preliminary Catalog, the aperture magnitudes from the ascending and descending epochs were averaged.
7.
In the NERSC version of the CatWISE2020 data products, the standard aperture magnitude uncertainties, w1sigm and w2sigm, are often 0.0 or 1.0 because of an error in the CatWISE2020 pipeline when averaging fluxes. The corresponding magnitudes, w1mag and w2mag, are correct, as are the individual aperture magnitude uncertainties. Users who are interested in standard aperture magnitudes should use the w1sigm_2 and w2sigm_2 for the uncertainties, as those are identical to what the w1sigm and w2sigm should have been. In the IRSA version, this error was corrected on 2020 August 25, and w1sigm and w2sigm values retrieved after this date are correct.
8.
A floor of 1 mas yr⁻¹ was imposed on the tabulated motion uncertainties for the CatWISE2020 data products, while for the CatWISE Preliminary data products the floor was 10 mas yr⁻¹. The lower CatWISE2020 uncertainty floor is now significantly smaller than the minimum measured scatter with respect to Gaia motion, as illustrated in the right column of Figures 21–24.
9.
CatWISE2020 source designations (source_name) include a lowercase letter suffix to distinguish sources that would otherwise have the same designation. In the CatWISE Preliminary Catalog, the first such source (usually the brightest) did not receive a suffix while subsequent such sources did, beginning with "b." In the CatWISE2020 data products, the first such source receives an "a" suffix.
10.
Sources in the updated unWISE catalog detection list may be omitted from the CatWISE2020 Reject Table if they are too near tile edges, probably due to the fitting region used by WPHOT being truncated. This does not affect the CatWISE Preliminary Reject Table because the MDET software does not find sources this close to tile edges.
11.
As discussed in Section 2.2, resolved galaxies result in multiple detections by crowdsource and hence multiple entries in the CatWISE2020 Catalog. Users interested in local overdensities in the galaxy distribution (such as for clustering analyses or detection of galaxy clusters) should be aware of this feature and take precautions to avoid misinterpretation.

**Figure 24.** Same as Figure 21 but for the Galactic Center tile (2657m288).
Download figure:
Standard image High-resolution image

The CatWISE2020 Catalog

Article metrics

Permissions

Author e-mails

Author affiliations

ORCID iDs

Dates

Abstract

1. Introduction