Every claim in the Short, and the sentence behind it
This patch of ocean glowed so bright that a satellite's night camera caught it from space.
"On the night of 25 July, the DNB detected a luminous anomaly south of Surakarta, Java, near 9.5° S, 111° E." "The detection amidst moderate moonlight conditions suggested a particularly strong source of emission."
Results, "Java, July-September 2019". The DNB is the Day/Night Band, "a new-generation spaceborne low-light imager" (Abstract).
On the early moonless nights, the patch was about the size of Iceland.
"Imagery on subsequent moonless nights confirmed that the initial detection was in fact part of a much larger milky sea" "approximately the same size as Iceland." "Approximate centroid locations and spatial area estimates correspond to the maximum observed size of each event."
Results, "Java, July-September 2019"; Table 1 note. These are excerpts from the size sentence and the table note. The comparison does not describe its size throughout the event.
The authors say it lasted at least 45 nights. The satellite lost sight of it in the middle.
"The longevity of this event, which lasted for at least 45 nights, by far eclipsed all other cases encountered in this study" "After losing sight of the luminous body on 10 August due to moonlight contamination, the DNB recaptured it on 25 August and tracked it for 2 weeks thereafter"
Results, "Java, July-September 2019". The 45-night span includes 10-24 August, a 15-night gap. Continuous glow during that gap was inferred, not directly observed by the satellite.
It's called a milky sea, and scientists think the light comes from bacteria.
"Milky seas are a rare form of marine bioluminescence where the nocturnal ocean surface produces a widespread, uniform and steady whitish glow." "Milky seas are understood as a kind of marine bioluminescence arising from a saprophytic relationship between luminous bacteria and microalgae that expresses on the macroscale."
Abstract; Introduction (reference numbers removed).
A few bacteria can emit detectable light. Typically, dense populations switch on and glow steadily.
"Even a few bacteria can produce light detectable to highly sensitive photometers, but typically they do not begin to emit such light until they reach critical populations of ~ 10⁸ cells mL⁻¹; then they turn on rapidly (within hours) and emit steadily." "This critical population is a proxy for autoinducer reaching concentrations sufficient to trigger bacterial quorum sensing."
Discussion, "A natural flask?" (reference numbers removed). "Counting its neighbours" is the Short's plain-English picture of quorum sensing. The paper calls the signal a "bacteria-secreted autoinducer" (Introduction). The Short's line "But they don't light up alone" is too absolute.
Scale that up, and a patch of sea glows.
"the visual effect is a steadily and uniformly glowing sea."
Introduction. This sentence describes the same process occurring "en masse", across "thousands of square km of ocean surface". It is the proposed bacterial explanation.
Sailors compared milky seas to a snowfield stretching to every horizon.
"Mariners have compared their appearance to a daylit snowfield that extends to all horizons."
Abstract. This describes mariner accounts of milky seas generally, rather than a ship sighting of the 2019 Java patch.
Ships reported it for over 200 years. Now satellites can find it.
"ship reports (Supplementary Table S1) and satellite detections of milky seas." "These findings show how spaceborne assets can now help guide research vessels toward active milky seas to learn more about them."
Figure 7 caption; Abstract. The caption dates the compiled ship reports to 1796-2010. The satellite's limits include the moonlit gap described above.
What the study doesn't say
It doesn't establish that an astronaut could see the glow with the naked eye from orbit. The authors say portions of the Java event "may have appeared visually bright to dark-adapted human vision" (Results, "Java, July-September 2019"). The Methods warn that the sensor-to-eye comparison "does not provide a direct translation between the appearance of DNB imagery and the visual perception of a milky sea". The Introduction says the sensor is "carried on two National Oceanic and Atmospheric Administration (NOAA) operational satellites".
The bacterial explanation remains a working hypothesis for this event. The main text reports no water samples from the 2019 patch. It says "The lone research vessel encounter happened by chance in the Arabian Sea, east of Socotra, in July 1985" (Introduction, reference number removed). The authors say testing their hypotheses requires the ability to "sample milky seas in situ across space and time" (Discussion, "The path forward"). The Short says "scientists think" and names no species.
The bacteria figures are modelled estimates. The Methods say "the estimates of total population assume that bacteria reside in a 1 cm-thick biofilm of surface water" and that "these total cell population estimates should be considered as very rough." The authors challenge the surface-slick idea: it "cannot explain many mariner accounts" (Discussion, "A natural flask?"). They propose a "natural flask", an isolated volume of water, as an alternative: "The flask hypothesis would reconcile cases where the surface-slick idea breaks down."
A satellite saw one before this paper. The Introduction describes how, "In January 1995", the S.S. Lima "traversed a milky sea" that "was later confirmed from space" by the "Operational Linescan System (OLS)" (reference numbers removed). It calls that sensor's imagery "poor quality". This paper demonstrates what the newer sensor can detect: "Here, we show initial examples of DNB-detected milky seas" (Abstract).
The real glow is described as whitish: "a widespread, uniform and steady whitish glow" (Abstract). The Short draws it that way. The patch's exact outline is an illustration. The paper's satellite images of the 2019 event are in Figures 4-5.
The source
Miller SD, Haddock SHD, Straka WC III, Seaman CJ, Combs CL, Wang M, Shi W, Nam S (2021). "Honing in on bioluminescent milky seas from space." Scientific Reports 11: 15443.
DOI: https://doi.org/10.1038/s41598-021-94823-z - free to read
The Short
Embed: https://www.youtube-nocookie.com/embed/SFleWbtV9tk
38 seconds, and it loops. Watch it on YouTube: https://youtu.be/SFleWbtV9tk
The numbers in the Short come from the paper: the Iceland comparison, 45 nights, and ship reports from 1796 to 2010. The Iceland outline is drawn at the same scale as the glowing patch. The correction below qualifies the bacteria line and the size and duration claims.
Correction - 2026-10-03
The Short says "But they don't light up alone". The paper says even a few bacteria can emit detectable light; dense populations typically switch to a steady glow. This page now keeps that qualification, identifies the moonlit gap within the authors' 45-night span, and ties the Iceland comparison to the early observations. It also corrects the description of the sensor's contribution and explains the authors' challenge to the surface-slick model.
Published 2026-10-02. Revised copy dated 2026-10-03; publication of this correction is pending.