Nocturnal Lights

Mirages at night

People are so used to associating (inferior) mirages with the desert and the heat of the day that they often overlook the possibility of mirages after dark. Yet the radiative cooling of the ground surface after sunset typically leads to near-critical inverted lapse rates at night. Particularly in dry climates, where the radiative cooling is hardly impeded by water-vapor absorption, the nocturnal boundary layer quite often has a steep enough inversion to produce superior mirages. Indeed, these have been used routinely by the Arabs as a navigational aid in the desert; the early-morning mirages are said to provide a view of the country “some 50 kilometres farther on”. (For comparison, the distance to the horizon under normal circumstances for a person standing on a smooth, level surface is about 5 km.)

Even in wetter climates, looming is normal at night, so that objects normally beyond the horizon in the daytime will routinely become visible at night if artificial light is available near them. This effect was used extensively in the 19th-Century Survey of India to extend baselines (in some cases, beyond 20 miles [32 km]).

At sea, or over lakes and rivers, the great thermal inertia of water allows inferior mirages to persist after sunset. Once again, quite distant land may be seen miraged, particularly hills or mountains well above the water.

Lights at night

The large dynamic range of the human eye allows lights to be seen at surprisingly great ranges at night. Once you are well away from the light pollution of civilization, so that full dark adaptation is possible, a single light bulb can be seen at a distance of many kilometers. (If it weren't for atmospheric extinction, an ordinary 100-W light bulb would be barely visible to a dark-adapted eye at about 100 km distance.) The lamps used in street lights are much brighter than this; and of course lights with directive reflectors (like automobile headlamps) also can be seen at much greater ranges than isotropic radiators of the same total power.

Typical car headlamps produce about 1000 lumens — comparable to the total light output of a 60-W incandescent light bulb — but of course most of this is concentrated into a fairly narrow beam by the reflector and lens optics. The output beam of a car headlight covers roughly 0.1 steradian, so the concentration by the optics is on the order of a thousand. That means a car's headlights can be seen about 30 times as far away as an ordinary light bulb of the same power. If they're pointing right toward you, and the air is clear, you should have no trouble seeing car headlights a couple of hundred kilometers away on a dark night.

Furthermore, once the light is trapped in a duct, its irradiance falls off only inversely with the distance (apart from extinction), instead of with the usual inverse-square law; of course radiance is preserved, but the solid angle subtended by the magnified and multiple images of the source increases with distance to make up the difference. So it isn't hard to see artificial lights at distances exceeding 100 km, if the air is very clear.

Photographic illustration

Ed Darack has photographed several fine examples of such miraged lights at night. He saw lights at the Otero County Prison on U.S. Route 54 north of Newman, NM, from a position some 50 miles (80 km) north of there, just northeast of the White Sands Gate, near Hurtz Spring. Some of his pictures were published in the May/June, 2008, issue of Weatherwise.

Historical examples

Even in the days before electric lights, surveyors used signal fires with reflectors behind them to measure long baselines at night. In establishing the metric system (which was originally based on the size of the Earth) in 1806, Biot and Arago used such a signal on the island of Ibiza that was observed from Desierto de las Palmas on the Spanish mainland, a distance of some 160 km. After two months of futile efforts, “We perceived in the field of the telescope a luminous point, very small, almost imperceptible, like a star of the fifth or sixth magnitude, but which was distinguished from a star by its immobility.” Multiple (ducted) images, some split into red and green components by atmospheric dispersion, were repeatedly seen during their observations. They noticed that the lights appeared brighter when the mirages multiplied the images.

A similar observation had been made by Joseph Huddart in 1797. He noticed that a lighthouse lamp appeared brighter when magnified by the inferior mirage.

Seeing lights at night

So it should hardly be surprising that distant lights are often seen unexpectedly at night — particularly in arid, flat places. In the mountains, there are often nocturnal mirages produced by drainage of cold air into valleys, so that observers below the resulting inversion see lights at distant settlements that are normally hidden by intervening low ridges. Observers on hillsides just above the inversion may see mock mirages of lights on the opposite side of a valley. Two, three, or more images of a single source may be visible; some may show colors (red and green being most common).

At ranges of many kilometers, the image of a small light source is blurred and enlarged by turbulence in the intervening atmosphere. This blurring is only a few seconds of arc at distances of a few km, but can become a minute of arc or more at great ranges — large enough to give the light a perceptible angular diameter, even to the naked eye.

People easily become disoriented in the dark, and imagine that a perfectly stationary light is “moving” or even “jumping” around. I've had students tell me they saw stars moving in the sky, during constellation-teaching sessions with undergraduates. (These were identifiable “fixed” stars, not artificial satellites.)

The distorted or multiple image of a distant miraged light is likely to be unidentifiable. Observers unfamiliar with such phenomena often attribute supernatural properties to the “mysterious” images.

In areas where these phenomena are common, they have received local names.


Marfa lights (West Texas)

For a sensible discussion of these miraged car headlights, see here. (Ignore the credulous stuff on Wikipedia; it does, however, have links to various other similar phenomena around the world.) The flat, arid area where the lights are seen is ideal for nocturnal mirage production. (I'm moderately familiar with the area, having worked at nearby McDonald Observatory back in the 1960s.)

There's an interesting experiment done in 2004 by the Dallas chapter of the Society of Physics Students, indicating that the lights are just miraged automobile headlights. A followup to this by Michael Hall in the June, 2006, issue of Texas Monthly agrees with them, and basically debunks the claims of pre-automotive observations.

And I've already referred to Ed Darack's 2008 article in Weatherwise on the Marfa Lights.

the Min Min lights (Australia)

A nice discussion of these mirages was published by John Pettigrew in Clinical and Experimental Optometry 86, No.2, 109–120 (2003). It's conveniently available on the Web at There are beautiful color photographs of the fine “Fata Morgana” mirages seen at dawn after a night in which the lights were seen, and a good discussion of the perceptual problems experienced by observers of nocturnal lights.

Siranui lights (Japan)

I have only some e-mail from a Japanese correspondent about these, but they seem to fall in the same general category.

Irrlichter (Germany)

This is the “will o' the wisp”, which supposedly misleads unwary travelers. The term is evidently a catch-all that subsumes many different phenomena, including glow-worms, phosphorescent fungi, and maybe even the burning swamp-gas that gets the blame all too often. Some fraction of the reports are probably nocturnal mirages, as well.

Wikipedia's article on “Will-o'-the-wisp” lists many more such phenomena.

General comments

Unidentified (and usually unidentifiable) lights occasionally seen at night are not very interesting to me. My purpose here is to point out that nocturnal mirages are much commoner than is generally supposed; and that lights — especially, miraged lights — can be seen at night to much greater distances than most people imagine. The combination of these two facts means that at least some of the nocturnal lights are miraged images of ordinary light sources at considerable distances from the observers.

In particular, there are sites on the Web that claim automobile headlights can't be seen beyond distances as short as four miles. Rubbish! Let me refer them to NOAA's website that tells some of the history of surveying — — in particular, the part that discusses the lights used in surveying by the U.S. Coast and Geodetic Survey:

Around 1900, bicycle acetylene lamps were used by the C&GS for the survey along the 98th meridian and were found useful for surveying distances up to about 34 miles. Larger acetylene automobile headlights were also eventually used successfully. Two of these automobile lamps stacked one above the other were observed at a distance of 133.9 miles at Pilot Peak, Nevada.

The development of the electric automobile headlight resulted in an excellent signal lamp. A bulb with a special filament was made at the request of the C&GS. In 1920, the light from two of these headlights was visible at a distance of 152.9 miles to an unaided eye (no telescope). By the mid-1920s, the electric signal lamp was the standard. …

… The signals described above allowed accurate triangulation methods to be used to survey the nation.

For readers accustomed to international units, 34 miles is about 55 km; 133.9 miles is over 215 km; and 152.9 miles is just over 246 km.

Modern headlights are considerably brighter than those used in 1920 — not to mention the older acetylene lights, which are compared with the electric ones in

E. G. Fischer
“The new triangulation signal lamp of the U. S. Coast and Geodetic Survey”
Scientific American 114, pp. 454, 456 (April 29, 1916)
as shown above the line in the table below:

Lamp type Electrical
Apparent Candlepower
at 100 feet
Luminous intensity (cd)
Acetylene   1 500 1.6
Ordinary headlamp 10.8 W  50 000 54
Special triangulation lamp 15 W 250 000 270

Regular sealed-beam 35 W   2 500
Halogen sealed-beam 35 W  27 000
Brightest legal headlight 150 000

I've added modern automobile headlights below the horizontal lines in the table. Even ordinary low-beam lights are an order of magnitude brighter than the special lamps used by the Survey after 1920 — those special lamps that were visible to the naked eye at nearly 250 km.

Of course, when the source of the lights is obvious, there's nothing mysterious about them. For example, look at the mirages over Lake Ontario shown on Kerry-Ann Lecky Hepburn's web pages. Notice how much easier it is to see the lights at night than the corresponding buildings in the hazy daytime pictures, at a range of 53 km. Put those lights twice as far away, and they'd still be easy to see at night; but nothing would be visible in the daytime.

When the air is cold, dry, and clear, nocturnal mirages should always be considered as a likely explanation for “unexplained” lights seen near the horizon.


Copyright © 2008, 2012, 2014, 2016 Andrew T. Young

Back to the . . .
main mirage page

or the GF home page

or the website overview page