On the night of 25 January 1849, Captain Kempthorne of the ship Moozuffer recorded in his log that the Arabian Sea around him had begun to glow. The phenomenon went on for hours. The ocean to every horizon had turned the colour of milk, or of liquid mercury. There was no moon, no obvious source of the light, and no apparent disturbance of the water — the sea around the ship was unusually calm, almost glassy, while everything above the surface and below it was apparently normal. Kempthorne, an experienced mariner familiar with the more common forms of marine bioluminescence (the blue-green sparkle that boats sometimes leave in their wake when disturbing dinoflagellates), recognised that this was something different. He wrote in his log that it looked “as if we were sailing over a boundless plain of snow, or a sea of quicksilver.” His description joined a small but consistent body of similar reports stretching back to roughly the year 1600 and continuing through the present day: sailors crossing certain remote stretches of the Indian Ocean, encountering an event they could only describe with the awkward word “milky,” and recording it as one of the strangest experiences of their working lives.
According to Colorado State University’s announcement of the new milky seas database, only about 400 milky-sea sightings have been credibly recorded across all 400 years of the historical archive. Only one photograph of the phenomenon is known to exist. Only one research vessel has ever sampled a milky sea directly while it was happening — the R/V Lim, in 1985. The combination of extreme rarity, remote location (the great majority of sightings occur in stretches of the Indian Ocean far from regular shipping lanes), and the absence of any reliable way to predict when a milky sea would happen has made the phenomenon nearly impossible to study by conventional means. For most of the past four centuries, milky seas were known only through the testimony of mariners, recorded in ships’ logs that were rarely read by scientists, and treated by the scientific establishment as either folklore or unconfirmed curiosity.
How satellites changed the picture
The transformation of milky seas from sailor’s tale to scientifically tractable phenomenon began in the mid-2000s, when Steven Miller — an atmospheric scientist at Colorado State University, and co-author of the 2025 database paper — realised that the new generation of weather satellites being launched at the time carried sensors sensitive enough to detect very low levels of nighttime visible light. The NOAA Day/Night Band sensor on satellites like Suomi NPP and NOAA-20 was designed to map city lights and to capture the faint glow of moonlit clouds, but Miller proposed that it should also be able to detect the steady, broad-area glow of a milky sea from orbit, if one happened to occur during a satellite overpass on a moonless night. The first satellite image confirming a milky sea was captured in the Java Sea around 2005, showing an event roughly 38,000 square feet in extent. A 2019 event south of Java covered approximately 100,000 square kilometres — comparable in area to Iceland — and remained visible from orbit across multiple consecutive nights.
As reported by National Geographic’s coverage of the database paper and its underlying historical work, the satellite record allowed researchers to confirm that the events sailors had been describing for 400 years actually existed and were not, as some sceptics had suggested, optical illusions or exaggerated reports of more conventional bioluminescence. The events were real, they were large, they could last for weeks, and they appeared with statistical preference in the Arabian Sea, the Java Sea, and the northwestern Indian Ocean — exactly the regions where the historical sailor reports had concentrated. The two records — four centuries of mariner accounts and two decades of satellite imagery — were describing the same phenomenon.
What is probably causing the glow
The biological identity of the glow remains, technically, unconfirmed. The leading hypothesis, supported by the single direct water sample taken by the R/V Lim in 1985, is that milky seas are caused by extraordinarily dense colonies of a luminous marine bacterium called Vibrio harveyi, living in association with surface algal blooms. The 1985 sample contained both the bacteria and an algae species that secretes a mucus capable of calming the ocean surface — which would help explain the eerily flat seas that sailors consistently describe in milky-sea encounters. The mechanism by which the bacteria coordinate their luminescence is one of the better-studied features of microbial biology and is called quorum sensing: bacterial populations release a chemical signal whose concentration in surrounding water increases as the population grows. When the chemical concentration crosses a threshold, the bacteria detect that they are in a sufficiently dense colony to make coordinated light production worthwhile, and a genetic switch turns the luminescence on across the entire population almost simultaneously.
The biological purpose of the luminescence, in this model, is not certain but is plausibly an evolved bait strategy. A glowing patch of ocean is visible to fish from substantial distances. Fish are drawn toward the light, consume the glowing algae, and inadvertently carry the bacteria into their gut — where the bacteria can survive, multiply, and eventually be excreted back into the water in a new location. The bacteria gain dispersal and a more nutrient-rich environment than the open ocean offers. The fish, presumably, gain very little. The arrangement, if confirmed, would be one of the larger examples of bacterial signalling translating into a phenomenon visible from low Earth orbit.
What the new database reveals
Per Euronews’s coverage of the 2025 Hudson and Miller paper, the combined historical and satellite record has now allowed researchers to identify statistical correlations between milky-sea events and broader climate patterns. The 2025 paper specifically links milky-sea occurrence to the Indian Ocean Dipole — a temperature-difference pattern between the western and eastern Indian Ocean that influences monsoon strength and rainfall across the region — and to the El Niño Southern Oscillation in the Pacific. The correlations are statistical rather than mechanistic, but they suggest that milky seas are not random occurrences. They appear preferentially during specific climate phases, in specific regions, under specific conditions of ocean temperature, surface circulation, and nutrient availability.
The practical implication is that the milky-sea database may eventually allow researchers to predict when and where a milky sea will occur with enough confidence to dispatch a research vessel in time to sample one. The single direct sample available to date — from 1985, more than 40 years ago — has been the entire empirical basis for the bacterial-bloom hypothesis. As noted by a Phys.org summary of the Hudson and Miller paper, getting one or two additional samples, taken from a milky sea while it is actively glowing, would allow molecular biologists to confirm the bacterial identity, characterise the algal partner, measure the chemistry of the bloom, and test the quorum-sensing mechanism directly. None of this is currently possible because milky seas are too rare and too remote to encounter on purpose. The database makes the encounter substantially more likely. The phenomenon that sailors have been describing for four hundred years may, within the next few decades, finally be examined under a microscope.
