Sky Survey · NEOWISE 2014–2024 + SPHEREx 2025–26 · automated search
Billions of infrared detections. Which ones move?
NASA's NEOWISE telescope photographed every part of the sky about every six months for ten years. We cut the sky into 12,288 equal tiles and, tile by tile, a program links each source's detections across the years, measures how far and how steadily it moves, throws out artefacts, and checks ESA's Gaia catalogue: a mover that Gaia already knows is explained. Then SPHEREx, which mapped the whole sky again in 2025–26, gets a say: the program reads SPHEREx's all-sky spectral cube at the spot where each unexplained mover should be by then, checks that it is there (and gone from where it was in 2014), and takes its 0.75–5 µm spectrum, looking for the methane dip and red colour of a cold brown dwarf. Whatever is left is ranked, so people can look at the most promising candidates first. A candidate is a lead, never a discovery.
NEOWISE measures the motion (2014–2024). SPHEREx has no source catalogue yet, so the program measures its DR1 cube directly; its ~13″ pixels confirm only fast motions (about 1″ a year or more) and can blend close neighbours.
- Detections searched
- —
- Sources linked across years
- —
- Sky searched
- —
- Movers found
- —
- Unexplained by Gaia
- —
- Confirmed by SPHEREx
- —
- Computer time
- —
Where we looked
Every tile searched, on one map of the whole sky
Each patch is one HEALPix tile of about 3.4 square degrees. Blue tiles hold movers that Gaia explains; amber tiles hold at least one unexplained candidate. Equatorial coordinates, east to the left.
- Searched, no movers
- Movers, all explained by Gaia
- At least one unexplained
- Not searched
- Milky Way plane
- Ecliptic
Loading…
Does it work?
We tested the search before trusting it
Two checks: fake movers hidden among the detections (how many does it find?), and famous fast stars it was not told about (does it find them on its own, with the right speed?).
Loading…
Ranked for people to investigate
The candidates
Highest score first. The score rewards no Gaia counterpart, a clean straight track over many years and a red colour (cold brown dwarfs are red in W1−W2), and penalises the opposite. Open a candidate to see its track and check it yourself.
High-confidence keeps movers seen in at least 12 visits whose motion is measured at 10σ or better, and hides those at a tile edge, in a crowded spot, saturated, or classed as likely artefacts. Red is a hint, not proof: cold brown dwarfs are red in W1−W2 because methane absorbs their light in W1 (3.4 µm). Filters and sort are kept in the address, so a shared link shows the same list.
Honest limits
What this search cannot tell you
- "Unexplained" means only that Gaia DR3 has no matching moving star. Many such objects are already known from other surveys (brown dwarfs are too faint and cold for Gaia), and some are artefacts the cuts missed.
- Only people looking at the images, then follow-up observations, can confirm anything new.
The run's own description of its method
Data and tools
- NEOWISE-R single-exposure source table
- NASA/JPL-Caltech/IPAC, read from NASA IRSA's open data on AWS (2014–2024)
- Gaia DR3
- ESA/Gaia/DPAC, queried through the ESA Gaia archive, or its mirror at ARI Heidelberg when the archive is unreachable
- SPHEREx and AllWISE sky maps (HiPS)
- NASA/IPAC Infrared Science Archive (IRSA)
- Aladin Lite
- CDS, Strasbourg Astronomical Observatory
- HEALPix tiles
- Górski et al. 2005, ApJ 622, 759