Updated explainer — written by an AI model
This account was written automatically by claude-sonnet-5-5 (Anthropic) from the official reports in this thread, with web search for background. It may contain mistakes; the reports themselves are the authority.

What happened

On 2026-10-15 at 13:15:36 UTC, the Einstein Probe (EP) satellite saw a brief flash of soft X-rays, now named EP261015a. EP is a space X-ray observatory. The flash lasted about 130 seconds and showed two peaks. It did not trigger the satellite's onboard alarm. The EP team found it in the data and sent an alert by hand.

Telescopes then found a faint, slowly changing optical source inside the X-ray error circle. At first it faded. Then, around 3.7 days after the trigger, it appeared to brighten again. Its nature is still unsettled.

The science behind it

Fast X-ray transients are short, bright X-ray flashes. Several kinds of event can cause them:
- A gamma-ray burst (GRB) seen slightly off-axis, or a weaker "X-ray flash". Both come from jets launched when a massive star collapses.
- A supernova shock breaking out of its star.
- The merger of two neutron stars.

When a jet drives into surrounding gas, it produces an "afterglow" that fades smoothly, usually as a power law, brightness ∝ t^-α with α near 1. Some collapsing massive stars also produce a supernova, which adds a slow brightening bump days later. Earlier EP events have been linked to supernovae, so astronomers check for this.

How it was detected: EP's Wide-field X-ray Telescope (WXT) uses "lobster-eye" micropore optics to watch a very large patch of sky in 0.5–4 keV X-rays. A keV is a unit of photon energy, and these are soft X-rays. The Follow-up X-ray Telescope (FXT) is a narrower, more sensitive instrument that EP points at the source afterward.

What the X-ray numbers mean:
- Position: WXT gave R.A. 44.341°, Dec −13.365°, with a 3 arcminute radius at 90% confidence (C.L.). That means a 90% chance the true position is inside the circle.
- Spectrum: the WXT fit uses a power law with photon index 1.45 (−0.71, +0.88). The index describes how steeply the photon counts fall with energy. A lower number means a harder spectrum. Afterglows often have values near 2, but the uncertainty here is large.
- Absorption: the fit fixed the Galactic hydrogen column density at 3.5×10^20 cm^-2. This is the amount of gas along our line of sight that absorbs soft X-rays, and it is low. The "unabsorbed" flux corrects for it.
- Flux: 6.25 (−2.91/+3.90)×10^-10 erg/cm²/s in 0.5–4 keV. Multiplying by the 130 s duration gives a rough fluence of order 10^-7 erg/cm². This is my own crude estimate and it assumes a constant flux. No fluence was reported.
- Energy and distance: no redshift has been reported, so the true energy and distance are unknown.

The optical numbers use AB magnitudes. This is a brightness scale in which a larger number means a fainter source. For example, r ≈ 22.7 corresponds to roughly 4 microjanskys (my estimate). The r band covers red-orange light. Most of these measurements come from "difference imaging". An archival image of the field is subtracted from the new one, which leaves only what changed. None of the magnitudes are corrected for Galactic dust. The gas column above suggests the dust dimming is small, probably a few hundredths of a magnitude (my estimate).

The optical candidate sits at RA 02:57:19.83, Dec −13:21:17.3. I estimate it lies about 0.8 arcminutes from the WXT centre, well inside the error circle.

The difference-image r-band brightness over time:
- 13.9 hr: 22.64 ± 0.12 (NOT, after refinement). The Las Cumbres Observatory (LCO) measured 22.83 ± 0.16 at a similar time.
- 15.6–17 hr: 23.04 ± 0.19 (COLIBRÍ).
- 1.54 d: 23.18 ± 0.10 (NOT). The NOT team quotes a decay of t^-0.51±0.15, which is shallow. They stress that the significance of the fading cannot be reliably determined, because subtraction between different instruments carries hidden systematic errors.
- About 3.7–4.4 d: about 22.5–22.6 (COLIBRÍ, Wendelstein, AZT-20).
- 5.25 d: 22.30 ± 0.24 (DFOT, tentative). A same-day 2-m LCO measurement gave r = 22.26 ± 0.06 and g = 23.51 ± 0.22. It gave only limits in i and z (fainter than 22.34 and 22.14).

A caution on one number: COLIBRÍ's circular 45991 lists the rebrightening as r = 21.58 ± 0.09. Three later circulars quote the same measurement as 22.58 ± 0.09. They are probably correcting a typo, but I can't confirm which is right, so treat the size of the rebrightening as uncertain.

The g−r colour from the day-5 data is about 1.2 magnitudes, which is quite red. This is my own difference of two uncertain numbers and I wouldn't lean on it.

Why this one is notable

The thread was posted because it reached 11 GCN Circulars, the top-10% threshold for X-ray transients. It has since grown to 14 reports.

The scientific interest is that the picture is ambiguous. The FXT, observing at 15.26 hours, found no X-ray source at the optical position. Its upper limit is about 1.33×10^-13 erg/cm²/s (0.5–10 keV, 90% C.L.). The WXT flux was thousands of times higher, although in a slightly different energy band and at an earlier time. A fast fade is expected, so this is not surprising.

FXT did detect an uncatalogued source elsewhere in the WXT circle, with a 10 arcsecond error radius. I estimate it lies about 2.3 arcminutes from the optical candidate. The NOT team notes it is probably the counterpart of a known quasar (QSO), a distant galaxy with an active nucleus, so it is likely unrelated. If so, the optical candidate is the best remaining link to the flash.

Complications:
- The candidate sits on a faint extended source, probably a galaxy. Legacy Survey magnitudes are g = 23.09, r = 22.39, i = 22.30 and z = 21.92. The galaxy is about as bright as the transient, so imperfect subtraction could produce a false signal.
- Detections from several telescopes, including SVOM/VT, DFOT, AZT-20 and MuSCAT3, argue that something real is there.
- A Terskol 2-m limit of 22.8 at 2.33 d is not constraining, because it is no fainter than the source.
- The WINTER near-infrared limit (J > 19 AB at 16.5 hr) is too shallow to say anything.
- COLIBRÍ suggested the late brightening could be a rising supernova. The NOT team says an unrelated supernova can't be excluded.

What to watch for next

- A spectrum of the optical source. This would give a redshift, which would allow a true distance and energy to be calculated. It would also show whether supernova features are present.
- Continued multicolour photometry. A supernova associated with a collapsing massive star should rise, peak and fade over weeks. An unrelated transient could behave differently, and the colour evolution would help sort these out.
- Late-time imaging after the transient fades. This would give a clean host-galaxy template and settle the subtraction problem.
- Radio or additional X-ray observations. Any later detection at the optical position would strengthen the link to the flash.
- Resolving the 21.58 versus 22.58 discrepancy in the COLIBRÍ report.

Further reading

- Einstein Probe mission overview, Chinese Academy of Sciences and ESA mission pages
- NASA GCN circular archive: gcn.nasa.gov/circulars (numbers 45955–46009 for this event)
- NASA HEASARC and Swift pages on gamma-ray burst afterglows
- Review articles on fast X-ray transients and X-ray flashes (search arXiv)
- Legacy Survey viewer: legacysurvey.org/viewer

Source reports
- EP261015a — X-ray Transient, Optical Transient — EP261015a — X-ray Transient, Optical Transient – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: Einstein Probe detection of a fast X-ray transient — EP261015a: Einstein Probe detection of a fast X-ray transient – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: COLIBRÍ optical observations — EP261015a: COLIBRÍ optical observations – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: SVOM/VT optical observation — EP261015a: SVOM/VT optical observation – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: J-band upper limit with WINTER — EP261015a: J-band upper limit with WINTER – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: refined analysis of EP-FXT observations — EP261015a: refined analysis of EP-FXT observations – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: Tentative fading of the candidate optical counterpart from the NOT — EP261015a: Tentative fading of the candidate optical counterpart from the NOT – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: COLIBRÍ optical observations of a rebrightening — EP261015a: COLIBRÍ optical observations of a rebrightening – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: Terskol Zeiss-2000 optical upper limit — EP261015a: Terskol Zeiss-2000 optical upper limit – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: Follow-up with the 1.3-m DFOT — EP261015a: Follow-up with the 1.3-m DFOT – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: Wendelstein 43 cm optical observations — EP261015a: Wendelstein 43 cm optical observations – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: AZT-20 Assy-Turgen optical observations — EP261015a: AZT-20 Assy-Turgen optical observations – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: Further optical observations with the 1.3m DFOT — EP261015a: Further optical observations with the 1.3m DFOT – Transient Alerts – HabitableZone (habitablezone.net)
- EP261015a: GSP multiband optical follow-up observations — EP261015a: GSP multiband optical follow-up observations – Transient Alerts – HabitableZone (habitablezone.net)