CatWISE Infrared Survey Catalogue

• CatWISE is an all-sky infrared catalogue derived from WISE and NEOWISE data that extends source detection limits and improves motion measurements.
• It employs epochal coadds and advanced detection algorithms, achieving deeper photometric sensitivity and enhanced astrometric precision compared to AllWISE.
• Applications include brown dwarf discovery, extragalactic source selection, and cosmological dipole studies, underscoring its broad utility in modern astrophysics.

Searching arXiv for CatWISE catalog papers and closely related applications. CatWISE is an all-sky infrared photometric and astrometric catalogue derived from combined WISE and NEOWISE survey data at 3.4 and 4.6 $\mu$m, corresponding to the W1 and W2 bands. Its defining purpose is to extend the WISE source census to fainter limits and to provide substantially improved multi-epoch motion measurements relative to AllWISE by exploiting a much longer temporal baseline. The CatWISE program first appeared publicly through the CatWISE Preliminary Catalog, which used data collected from 2010 to 2016 and reported 900,849,014 sources (Eisenhardt et al., 2019). The subsequent CatWISE2020 release expanded the time span to 2010 Jan. 7 through 2018 Dec. 13 and increased the catalogue to 1,890,715,640 sources over the entire sky (Marocco et al., 2020). Beyond its role as a reference catalogue, CatWISE has become a widely used infrastructure for brown dwarf discovery, extragalactic source selection, and large-scale anisotropy studies.

1. Origins, scope, and catalogue releases

CatWISE was designed as a successor to AllWISE for source selection and motion measurement in W1 and W2. The CatWISE Preliminary Catalog consists of 900,849,014 sources measured in data collected from 2010 to 2016, a dataset representing four times as many exposures and spanning over ten times as large a time baseline as that used for the AllWISE Catalog (Eisenhardt et al., 2019). This preliminary release adapted AllWISE software to measure sources in coadded images created from six-month subsets of the WISE and NEOWISE data, each representing one coverage of the inertial sky, or epoch, and included measured motion over 8 epochs across a 6.5 year span (Eisenhardt et al., 2019).

CatWISE2020 extended that framework. The CatWISE2020 Catalog consists of 1,890,715,640 sources over the entire sky selected from WISE and NEOWISE survey data at 3.4 and 4.6 $\mu$m collected from 2010 Jan. 7 to 2018 Dec. 13 (Marocco et al., 2020). Relative to AllWISE, this dataset adds two years to that used for the CatWISE Preliminary Catalog, bringing the total to six times as many exposures spanning over sixteen times as large a time baseline as the AllWISE catalog (Marocco et al., 2020). The catalogue is publicly available in the WISE/NEOWISE Enhanced and Contributed Products area of the NASA/IPAC Infrared Science Archive (Eisenhardt et al., 2019, Marocco et al., 2020).

The catalogue’s scientific niche is defined by the conjunction of all-sky coverage, W1/W2 sensitivity, and proper-motion capability. A plausible implication is that CatWISE occupies an intermediate regime between static all-sky source lists such as AllWISE and explicitly time-domain products such as unTimely, which tabulates independent detections on time-resolved unWISE coadds rather than unique merged sources (Meisner et al., 2022).

2. Construction methodology and measurement model

The CatWISE Preliminary Catalog uses the unWISE processing pipeline to create deep coadds from available exposures per sky location, then measures sources on six-month epochal coadds rather than on individual exposures (Eisenhardt et al., 2019). Detection in the preliminary release was performed simultaneously in W1 and W2 via the Multiband Detection software, MDET, using PSF-matched filtering on the full-depth unWISE coadds (Eisenhardt et al., 2019). Source properties, including positions, photometry, and motions, were then measured on eight separate epoch coadds (Eisenhardt et al., 2019).

The motion model is a least-squares linear fit to source position as a function of time:

$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$

where $\vec{x}(t)$ is the position at time $t$, $\vec{x}_0$ is the position at a reference epoch $t_0$, and $\vec{\mu}$ is the proper motion vector in mas yr$^{-1}$ (Eisenhardt et al., 2019). For most of the sky, ascending and descending scan directions were processed separately because of different PSF orientations, then merged using inverse-covariance weighting (Eisenhardt et al., 2019).

CatWISE2020 introduced a major methodological change in source detection. The detection list for the CatWISE2020 Catalog was generated using crowdsource rather than the detection software used for AllWISE and the CatWISE Preliminary Catalog (Marocco et al., 2020). This crowded-field PSF-fitting detection algorithm was identified in the CatWISE2020 paper as one of the two principal reasons, together with the expanded temporal baseline, that the newer release contains roughly twice as many sources as the preliminary release (Marocco et al., 2020). The same paper notes that CatWISE2020 can detect up to $5\times$ more sources in the Galactic plane than the preliminary catalogue (Marocco et al., 2020).

The catalogue architecture is therefore anchored in coadded-epoch source measurement with global source selection over the full survey time span. This suggests that CatWISE is optimized for accurate positions, motions, and integrated W1/W2 photometry of unique sources rather than for explicit per-epoch variability characterization. That contrast is made explicit by unTimely, which provides independent per-epoch detections and no cross-epoch source association in its release (Meisner et al., 2022).

3. Photometric depth, astrometric precision, and comparison with AllWISE

The CatWISE Preliminary Catalog achieved deeper photometric limits than AllWISE. From comparison to Spitzer, the SNR$\mu$0 limits in magnitudes in the Vega system are W1$\mu$1 and W2$\mu$2, compared to W1$\mu$3 and W2$\mu$4 for AllWISE (Eisenhardt et al., 2019). The same comparison showed that CatWISE positions have typical accuracies of 50 mas for stars at W1$\mu$5 mag and 275 mas for stars at W1$\mu$6 mag, while proper motions have typical accuracies of 10 mas yr$\mu$7 and 30 mas yr$\mu$8 for stars with these brightnesses (Eisenhardt et al., 2019).

CatWISE2020 further improved completeness depth. Its 90% completeness depth is W1$\mu$9 mag and W2$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$0 mag, which is 1.7 mag deeper than in the CatWISE Preliminary Catalog (Marocco et al., 2020). From comparison to Gaia, CatWISE2020 motions are accurate at the 20 mas yr$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$1 level for W1$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$2 mag sources, and at the $\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$3 mas yr$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$4 level for W1$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$5 mag sources; this is described as a 12$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$6 improvement over AllWISE (Marocco et al., 2020).

A concise comparison of the two principal public releases is useful.

Release	Time span	Sources	Key stated performance
CatWISE Preliminary	2010–2016	900,849,014	SNR$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$7: W1$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$8, W2$\vec{x}(t) = \vec{x}_0 + \vec{\mu}(t-t_0),$9
CatWISE2020	2010 Jan. 7–2018 Dec. 13	1,890,715,640	90% completeness: W1$\vec{x}(t)$0, W2$\vec{x}(t)$1

The catalogue is commonly compared with unTimely because both rely on unWISE products but serve different scientific use cases. In the COSMOS field, the unTimely summary gives approximate CatWISE2020 limits of $\vec{x}(t)$2W1$\vec{x}(t)$3 and $\vec{x}(t)$4W2$\vec{x}(t)$5 for 50% completeness, with unTimely going deeper and offering repeated epochal detections (Meisner et al., 2022). Since those values are explicitly marked approximate in the summary, they are best interpreted as field-specific comparative guidance rather than a replacement for the official CatWISE2020 depth characterization.

4. Relation to other WISE-derived catalogues and photometric products

CatWISE belongs to a wider ecosystem of WISE-based catalogues. AllWISE is the immediate predecessor and provided the shorter-baseline reference point that CatWISE sought to surpass in both depth and motion precision (Eisenhardt et al., 2019, Marocco et al., 2020). The unWISE coadds are part of CatWISE’s image-processing foundation, and The Tractor forced-photometry program illustrates a complementary strategy in which WISE fluxes are measured at prior optical positions rather than through blind WISE-only detection (Lang et al., 2014).

The distinction between CatWISE and forced-photometry catalogues is methodological. In the SDSS-based forced photometry of Lang, Hogg, and Schlegel, WISE fluxes were measured for over 400 million SDSS sources using measured SDSS source positions, star-galaxy separation, and galaxy profiles, on unWISE coadds with The Tractor (Lang et al., 2014). That approach yields fluxes for all prior optical sources, including many below the native WISE detection threshold (Lang et al., 2014). By contrast, CatWISE is a WISE/NEOWISE source catalogue selected from infrared data themselves and provides photometry and motions for those infrared-selected sources (Eisenhardt et al., 2019, Marocco et al., 2020).

The distinction between CatWISE and unTimely is equally important. unTimely presents a full-sky, time-domain unWISE catalogue based on time-resolved unWISE coadds spanning 2010 through 2020, with detections extracted independently on each biannual sky pass using crowdsource and low detection thresholds of S/N$\vec{x}(t)$6 in W1 and S/N$\vec{x}(t)$7 in W2 (Meisner et al., 2022). It tabulates 23.5 billion detections at W1 and 19.9 billion detections at W2, rather than unique merged objects (Meisner et al., 2022). The summary explicitly contrasts CatWISE as being based largely on long-baseline static coadds with only two effective epochs per source for many applications, whereas unTimely preserves approximately 16 epochs per sky location from 2010 to 2020 (Meisner et al., 2022). This suggests a practical division of labor: CatWISE for all-sky static-source and motion work; unTimely for epochal variability and fast-mover searches.

Some WISE-based science requires additional photometric caveats. For bright saturated AGB stars, systematic magnitude-dependent offsets exist in WISE W1 and W2 PSF-fit photometry, and empirical calibration formulae have been derived for saturated AGB sources using ISO spectra and DUSTY modeling (Lian et al., 2014). The paper explicitly warns that analyses based on WISE/AllWISE/CatWISE catalogue products without this specific correction will introduce systematic errors for saturated AGB stars (Lian et al., 2014). That issue is not specific to CatWISE, but it is directly relevant to CatWISE users because CatWISE inherits WISE-band photometric systematics for bright saturated objects.

5. Brown dwarfs and the solar-neighborhood discovery space

One of CatWISE’s most visible scientific impacts is in the discovery of faint, red, high-proper-motion brown dwarfs. Earlier WISE work had already established that ultracool dwarfs occupy distinctive color space because W1 overlaps strong methane absorption while W2 samples a relatively unobscured spectral region, making late-type T dwarfs red in W1$\vec{x}(t)$8W2 (Burgasser et al., 2011). WISE-based selections using W1$\vec{x}(t)$9W2 $t$0 mag enabled the discovery of late-type T dwarfs and later Y dwarfs in AllWISE-era work (Burgasser et al., 2011, Schneider et al., 2015).

CatWISE extended this discovery channel by leveraging a longer time baseline. In the Spitzer follow-up study of cold CatWISE discoveries, CatWISE is described as a joint analysis of archival WISE and NEOWISE data that improved upon AllWISE motion measurements by leveraging a $t$1 time baseline enhancement, from 0.5 years for AllWISE to 6.5 years for CatWISE (Meisner et al., 2019). That motion selection yielded a large sample of previously unrecognized brown dwarf candidates, many with archival detections exclusively in WISE W2, suggesting that they could be both exceptionally cold and nearby (Meisner et al., 2019). Among motion-confirmed discoveries, seventeen had a best-fit Spitzer [3.6]$t$2[4.5] color most consistent with spectral type Y (Meisner et al., 2019).

A concrete example is CWISEP J193518.59$t$3154620.3, discovered in the CatWISE catalog (Marocco et al., 2019). Follow-up Spitzer photometry yielded ch1$t$4ch2 $t$5 mag, and the inferred effective temperature was $t$6--$t$7 K with an estimated distance of 5.6--10.9 pc (Marocco et al., 2019). Combining WISE, NEOWISE, and Spitzer gave a proper motion of $t$8 mas yr$t$9 and $\vec{x}_0$0 mas yr$\vec{x}_0$1 (Marocco et al., 2019). These figures place CatWISE squarely in the observational regime of the coldest nearby substellar objects.

The methodological pattern across these brown-dwarf studies is consistent. CatWISE furnishes W1/W2 colors, faint-source sensitivity, and motion measurements; Spitzer follow-up then supplies deeper 3.6 and 4.5 $\vec{x}_0$2m colors for temperature and spectral-type inference (Meisner et al., 2019, Marocco et al., 2019). This suggests that CatWISE’s principal value for substellar work lies in candidate generation and motion confirmation rather than complete physical characterization.

6. Extragalactic applications and cosmological use

Although CatWISE is often associated with brown dwarf searches, its all-sky W1/W2 photometry and astrometry also support extragalactic source selection. Earlier AllWISE work demonstrated that mid-infrared colors can identify large AGN samples with low stellar leakage using W1, W2, and W3 color criteria (Secrest et al., 2015). CatWISE inherits the deeper W1/W2 basis for analogous infrared-selected source studies, especially when proper-motion filtering is useful for rejecting stars.

A more specific CatWISE application is the study of the cosmic dipole using infrared-selected quasars. In a Bayesian analysis confronting Planck with catalogues of cosmological sources, CatWISE2020 was used to define a sample of 1,621,329 quasars selected with $\vec{x}_0$3 and $\vec{x}_0$4, together with sky masking for the Galactic plane, bright sources, and poor-quality regions (Land-Strykowski et al., 23 Sep 2025). The source-count dipole model for each sky pixel was written as

$\vec{x}_0$5

with an additional ecliptic-latitude bias correction specific to CatWISE (Land-Strykowski et al., 23 Sep 2025).

In that analysis, the CatWISE dipole corresponded to an inferred velocity of $\vec{x}_0$6 km s$\vec{x}_0$7 with direction $\vec{x}_0$8 degrees (Land-Strykowski et al., 23 Sep 2025). The joint Planck–CatWISE comparison produced a Bayesian suspiciousness of $\vec{x}_0$9 and a quoted significance of $t_0$0, which the authors interpret as severe tension between CatWISE and Planck under the kinematic interpretation (Land-Strykowski et al., 23 Sep 2025). The same study reported strong concordance between CatWISE and NVSS and suggested that their dipoles may arise from a common astrophysical signal (Land-Strykowski et al., 23 Sep 2025). Whether that result reflects cosmological new physics, astrophysical anisotropy, or residual systematics is an open question in the cited work; the paper itself frames the result as a challenge to the standard model if it persists (Land-Strykowski et al., 23 Sep 2025).

CatWISE thus participates in two very different extragalactic regimes: object-level selection of quasars and large-scale statistical tests of isotropy. A plausible implication is that the catalogue’s combination of enormous source counts and uniform all-sky mid-infrared coverage is as consequential for precision cosmological systematics as it is for rare-object discovery.

7. Limitations, caveats, and scientific legacy

CatWISE has several operational caveats that shape its use. First, the catalogue’s astrometry derives from WISE rather than a Gaia-recalibrated frame in the basic processing; the unTimely summary describes CatWISE astrometry as also using WISE astrometry rather than being generally Gaia-recalibrated, with forced photometry used to measure epochs in that comparison context (Meisner et al., 2022). Second, large, resolved galaxies can be split into multiple sources in CatWISE2020, although the release incorporated special handling using HyperLEDA and unWISE bitmasks to mitigate this behavior in many cases (Marocco et al., 2020). Third, bright saturated sources in W1 and W2 can suffer systematic photometric biases, as shown for AGB stars (Lian et al., 2014).

There are also scope limitations. CatWISE is not a catalogue of extended-source fluxes; studies of Galactic bubbles based on WISE and Herschel explicitly contrast their extended-source catalogue with CatWISE, noting that CatWISE primarily lists unresolved sources (Bufano et al., 2017). Similarly, CatWISE is not a per-epoch time-domain detection database in the sense of unTimely (Meisner et al., 2022). These distinctions matter because they delimit what can be inferred directly from the published catalogue without image-level reanalysis.

Its scientific legacy is nevertheless broad. CatWISE2020 is described as the largest all-sky mid-infrared catalog to date, with proper motions, unprecedented faint sensitivity, and improved Galactic-plane coverage (Marocco et al., 2020). In brown-dwarf science it has enabled the identification of exceptionally cold solar-neighborhood objects (Meisner et al., 2019, Marocco et al., 2019). In survey methodology it established a bridge between static WISE catalogues and more explicitly time-domain products (Eisenhardt et al., 2019, Meisner et al., 2022). In cosmological source-count studies it has supplied one of the most statistically significant infrared-selected dipole measurements presently discussed in the literature (Land-Strykowski et al., 23 Sep 2025).

Taken together, these uses define CatWISE less as a single-purpose catalogue than as a foundational W1/W2 astrometric-photometric layer for modern all-sky infrared astronomy.