Sky Observations

Light and color, clouds and other physical observations by day and night

Contrail Shadows

On the sunny afternoon of 29 October, I noticed that the sky was full of east-west contrails, some fresh, some dissipated. There were high clouds in the northeast, but the sky was mainly clear except for a few thin misty clouds in the south. I saw that one of these clouds was cut in two by a definite band, and the blue sky could be seen in this separation. What could have caused this strange linear separation? There was a prominent thick white contrail to the north (as I saw it) of the dark band, and in exactly the same direction. This is shown in the photograph. The dark band was, indeed, the result of the shadow of the contrail on the thin cloud. The contrail was much higher than the cloud, and cut off the sunlight that was illuminating it. We usually see opaque clouds by reflected sunlight, but these thin white clouds were seen by the sunlight scattered in them, by transmission, something I had not recognized before. I now noticed that there was a second layer of thin cloud above the first, and it, too, showed the linear shadow of the contrail. When the sunlight is cut off, we can see the sky behind through the thin cloud, which is normally masked by the scattered light. Be sure you understand that we were not seeing the shadow of the contrail on a cloud, something quite ordinary, but an effect on the other side of the cloud, that made it look as if it really were in two parts. The thin cloud could be seen only when directly illuminated by the sun, and without this it became invisible. These clouds seem to be called translucidus. Contrails are composed of ice crystals, and appear at altitudes over 30,000 feet, where the water in the exhaust freezes quickly.

AWACS aircraft have been observed to leave large, looping circular contrails over the North Sea (and probably elsewhere) that may be the only clear evidence of human activity on the earth visible from space. See the website in the References for photographs of contrails.

Billow Clouds

Sometimes, high, sheet-like clouds are regularly striated in patches near their edges, or such striated regions exist independently. These seem to be called billow clouds, a name that does not seem sufficiently descriptive. I have long been fascinated by the question of what determines the "wavelength" of the striations, and indeed of a good explanation of their formation. Examples are described in F. H. Ludlam and R. S. Scorer, Cloud Study (London: John Murray, 1957), with some indication of the factors involved. It is said that they have not received extensive study.

The billows are said to be arranged at right angles to the "wind shear," but this probably simply means a change with height of wind in a fixed direction. Their wavelength is said to range from 50m to 1km, perhaps usually about 200m. In any particular case, the wavelength appears to me to be quite constant, and the billows are linear, not curved. They appear to form in water-droplet clouds, and, like all clouds, are a dynamic phenomenon. In this case, the wind blows through them, and they may contain upwards and downwards motion. To understand them, it is necessary to understand the oscillatory mechanism--the inertia and the restoring force. The oscillations are probably driven by the evaporation and condensation of water. I think they form on the leading edge, where the wind flows into the cloud, and will try to verify this by observation. This would mean that they form near the critical level for condensation, and may show rising and falling waves.

While observing the contrail shadows on 29 October (see above) billow clouds were apparent in the northeast, in altostratus or altocumulus, as a front was approaching, and the wind seemed to be southwesterly, but I did not take care to notice the winds at higher levels. On 1 November, there were fragments of clouds at all levels in the sky in the early afternoon, but no hint of billows anywhere. What are the conditions for their appearance?

Billow clouds seem to be attributed to wave instabilities at the interface between two sheets of air with relative velocity between them. This interface has to be a vortex sheet, with many interesting properties. Linear analysis shows an instability called the Kelvin-Helmholtz instability. A small wave disturbance grows exponentially when the relative velocity exceeds a certain limit. This simple theory does not give any account of amplitude or wavelength, however, which leaves the most interesting part of the question unanswered. The wave moves at the average of the velocities of the layers, which should be susceptible to observational verification. It is exciting to contemplate that observed billow clouds may always be due to this mechanism, and the wavelength is somehow determined by the relative velocity.

A rather extreme form of a billow cloud is the Morning Glory of the Gulf of Carpentaria in northern Australia, shown in the photograph on the right approaching Burketown from the northeast. This photograph is from the excellent web site Morning Glory, which contains full information and an explanation of the wave phenomenon.

Sun Dogs

There have been high cirrus clouds for the past couple of days, and I was hopeful of seeing some halo phenomena around the sun, but had little luck. The moon is near new, so it was out of position. I did not see anything, until a little after 3 pm on 22 November, when I noticed bright parhelia, or sun dogs. I managed to get a photograph of them, one on each side of the sun a little more than 22° away. The sun's altitude was about 12°, and there happened to be cloud in the right place. They were brownish on the side towards the sun, and bluish on the other. They are rather common in Denver, but this was a particularly good showing, and I hope the photo comes out well. I did not observe any 22° halo. My observing position is bad, discouraging easy views of the sky, so I perhaps miss more halos than I see.

Tricker's book says the parhelia are due to hexagonal plates falling with their breadth horizontal, and rays entering somewhat skew, so that their position is outside the 22° halo formed by refraction in tumbling hexagonal prisms. They correspond to minimum deviation, of course. They disappear when the sun's altitude is greater than 60°, but here the sun was low.

In the twilight just after sunset, I noticed pinkish-orange clouds in the west. In their midst was a bright, silvery contrail. Obviously, they and the contrail were not illuminated by the same light. The contrail was much higher (though it did not appear so) and still in bright sunlight.

At 4 pm the next evening, the sun was low and bright, the atmosphere unusually clear and haze-free. There were linear wisps of cirrus high in the sky, as well as some on the horizon, but no halo. A shred of cloud to the right was just in the right position, 22° from the sun, for the halo, and there was a brilliant column of light, fairly short, but very distinct, that was the sun dog merged with the halo, as happens for a low sun, where the line of view is close to a plane perpendicular to the refracting edges of the hexagonal plates of ice. The sun dog was quite brilliant, since most of the plates must have been aligned, not randomly distributed as for the halo. Perhaps the sun dog and related phenomena are due to these plates, the regular halo to long prisms. The higher the sun, the more it is separated from the 22° halo. I am looking for the arc of Lowitz, connecting the sun dog with the halo and tangent to the halo. Perhaps this arc is also due to the plates, as is suspected, though no good theory exists. I think it is due to variations from the predominantly horizontal attitude of the plates. See Bob Fosbury's site below.

Lunar Halo

The lunar halo is a rather common occurrence. It usually consists of the 22° halo alone, without all the various possibilities of the solar halo phenomena, since the light intensity is much less. Many people have not seen the halo, however, or recognize its significance, so it may be good to describe a typical occurrence in this place. Last night, the 9th December, I finally saw a lunar halo after looking for some time. I know this is a night-time phenomenon, but I think it is better treated here. Halos may not be as common in Denver as in some other places, but it is still mainly a matter of looking. I generally always check the sun in the morning and the nearly-full moon for the possibilities of a halo. What is required is a thin veil of cirrostratus, hardly noticeable by itself. Thick clouds or a crystal-clear sky preclude halos.

The sky contained scattered high-level clouds, and the upper winds were strong from the west. About 19.30 the moon showed a corona as it shone through high-level clouds, and I looked for halo traces. An arc was visible in a darker patch, but it took a while to recognize it distinctly. As I examined the clouds at about this distance from the moon, I detected about 75% of a complete circle, interrupted here and there, and of varying intensity. The halo was not brighter than the patchy illuminated clouds. At 20.00 the cloud was thickening, and the halo grew less distinct. When this thicker cloud rapidly blew away, it was replaced by a much thinner layer, and at 20.20 the halo was quite distinct against the darker sky and nearly complete. By 20.35, however, only about 40% of the halo, formed by the upper part against a darker sky, was visible, the lower part obscured by cloud. By 20.40, the sky was clearing rapidly, and the remaining halo weakened. At 21.10, the sky was completely clear, and there was no halo. The moon shone without a corona, and stars were evident.

The rapid changes should be noted, since they are typical. The halo is a cloud phenomenon, not a moon phenomenon, and depends greatly on the distribution of cloud. A full, 360° halo happens only by chance, when the sky is covered by a thin, uniform layer of ice-crystal cloud. I did not see halo phenomena in contrails, nor was there the slightest hint of moon dogs (the moon was actually too high for moon dogs). The 22° halo is at the position of minimum deviation for refraction at a 60° edge in ice crystals that are hexagonal prisms with random orientations. The crystals must be a certain minimum size for a good halo, and wave properties of light are probably important (though most students usually just use ray tracing). I did not think to investigate the polarization, but will the next time.

At midnight, 11-12 December 2000, after very cold air had arrived, and the temperature was near 0°F with snow on the ground, the sky was covered with cirrostratus quite uniformly, and the moon looked as if seen through frosted glass. A complete 22° halo was visible, not prominently but quite definitely, as a brightening. There were no solar halos observed at any time during this period, however.

Lunar Corona

From about 8.0 to 8.15 pm MST, 5 February 2001, a beautiful classic corona was seen about the 10-day old moon. High clouds were blowing rapidly westward, and no haze was evident, though the bright corona showed that there was a veil across the moon. The aureole was whitish, shaded with brown on the edge, about 7° in diameter, resembling a shaded sphere. The ring surrounding it was about 10° in diameter. The inside edge was blue, then a band of green, a band of yellow, and the outside edge was red. The colors were not saturated, but were quite definite, and merged into one another. The pattern was almost perfectly circular.

The corona is a diffraction phenomenon, giving colors in the reverse order to the colors produced by minimum deviation in the rainbow. The central portion, the aureole, is very commonly seen, but the colored rings are rarer. The short time this corona was visible emphasizes the transient nature of many atmospheric optical phenomena, and the necessity for keeping a good lookout. The corona is easy to distinguish from the very different halo phenomena, by its smaller size and brighter color.

A corona was observed at about 10 pm MST, 12 April 2003. The moon shone through thin cirrostratus, and a slight halo was present. The aureole was about 3° in diameter, brownish on the outside. There was a dark ring, and then a fainter ring of 6° diameter, bluish in color. Outside a second dark ring was a still fainter ring, making a total of 9° in diameter. Later, as the cloud thinned, the corona shrank in diameter, becoming not more than 1.5° in diameter, with indistinct outer rings. Incidentally, I noticed that the halo passed just inside Jupiter. The distance from the moon to Jupiter was 22.8°.

Twilight Colors

As I watched the sunset a few days ago, the clouds were bright orange, and the sky behind was blue, actually cyan. The clouds reflected the low sunlight depleted of blues, while the sky behind showed the blues that were scattered. The two colors were, apparently, complementary.

John Wood Dodge was a famous painter of miniatures in the mid-19th century, who liked to show an orange and blue sky in the background. He made the sky orange, and the clouds blue! It is not unusual for artists never really to see what is before them.

Stars By Day

Since the beginning of the year, I have been quite pleased to discover that there is more to the daytime sky than sun and clouds. Many know that the moon is usually visible by day when it is above the horizon, but few realize that Venus is also a daytime object. When the sun is up, the solar rays are scattered by the atmosphere above us, so that a blue veil is drawn across the sky when the air is clear. Anything we see beyond this has to be bright enough to compete with the veil. The moon, which does not really have a bright surface, does quite well in showing through. Its surface can be seen in binoculars by day without the dazzling light of nighttime. Venus, visual magnitude -4, has a much brighter surface, and easily punches through the veil, looking like a small hard diamond.

Nevertheless, it is difficult to locate Venus in the daytime sky. If I find her current position from my program Sidera, then look with binoculars, I can usually find her in a few minutes. The binoculars are not necessary for seeing her, for once found, she is easily visible without them. If I let my vision wander, it may be difficult to find her again, however. Even without binoculars, I have found Venus, with more or less difficulty. Binoculars simply make the finding easier.

At noon on 2 March 2001, the moon was in the east, and Venus was easily found high in the sky. I looked around the moon to see if Jupiter could be seen. I did not find Jupiter, but I did see a reddish spot north and west of the moon that was very peculiar. This spot moved slowly away to the northwest, and was apparently an orbital object, perhaps the International Space Station. The color was remarkable, and was probably the color reflected from solar panels, which would be reddish. This object was dimmer than Venus, perhaps magnitude -3, and it showed why such things are usually hard to see. I noticed I could look at a perfectly blank blue sky around the moon, and then this ruddy object would suddenly appear, as if just turned on. As long as I kept it in my visual field, it was constant and quite apparent. If I lost it, I had to search again, and again it would pop into view. It is apparent that objects of this kind must be recognized as something before they are perceived. It is not simply a matter of looking in the right place, but of the visual system recognizing that the object is really something. The brightness of the red object was just right to make this effect plain. Incidentally, by night the redness might not be as apparent as it was by day.

Jupiter is probably brighter than the blue sky, but is more difficult to perceive because it is dimmer, only magnitude -2. I plan to look for Jupiter, but the search will probably be inconclusive. A larger telescope may make an object of this type more easily visible by day. The reason it is actually not seen is, however, that it will not be recognized as a real object, not that it is dimmer than the blue sky. There is, of course, a limit to this, and the strongest telescope will not resolve dim objects against the background.

It is an old misconception that stars can be seen by day from the bottom of a well. It is not the light around the observer that matters, but the contrast with the sky background. The following paragraphs are reports of observing Venus by day from time to time in the past weeks.

At about 2.25 pm 2 January 2001, I noticed the first quarter moon easily visible, with the sun in the south west and a clear blue sky. Venus was supposed to be about halfway between the moon and the sun at this time, so I searched for it with binoculars. It was easy to find, and Venus was a brilliant spot. Without the binoculars, I searched for Venus, using the binocular sighting as a guide. It was not so easy to find, but was quite definitely visible once found, high in the sky near the meridian. The contrast with the sky background was not as great as in the binoculars. Both the moon and Venus were sufficiently brighter than the sky to be seen.

Since the above observation, I have found Venus several times without binoculars during the day. Today, which was bright and sunny, with a few high clouds, I looked in the southeast with binoculars after checking where she should be, and quickly found Venus. Venus, like the moon, is a daylight object! Venus, at magnitude -4, is the next brightest object after the sun and moon. Jupiter, and Mars at his brightest are magnitude -2, and may just be possible. Next comes Sirius, and Mercury at his brightest, at magnitude -1.5. Saturn is magnitude 0, so at this level several bright stars could be seen as well. Even if these objects cannot be seen in full sunlight, the magnitudes may show the order in which they will appear as the sky darkens, or will vanish as the sky brightens. Mars has a wide variation in brightness, from -2 to +1.5, since its distance from Earth varies greatly. Mercury has a great range, from -1.5, as bright as Sirius, to +4. When it is dimmest, we see only a thin crescent, since it is between us and the sun, like the new moon. Uranus is magnitude 6, just visible to the naked eye under good conditions. Neptune is +8, difficult for binoculars, and Pluto is an impossible +14.

On 25 February 2001, Venus was clearly visible in the sky well before sunset, but the two-day crescent moon was not. At sunset, about 4.45 pm, Venus was brilliant, and the narrow crescent had appeared. The moon was only 30° from the sun, so the crescent was only 2' wide, a fifteenth of the moon's diameter. It seems clear that Venus is much more reflective than the dark rocks of the moon, and so was seen while the surface brightness of the crescent was still obscured by skylight. I have seen Venus quite a few times by day, now that I have looked for her. The only problem is finding her in the light, not seeing her once located. This is good practice for seeing supernovas.

The Moon in Binoculars

In an August night, the clouds of the afternoon's storms had dissipated, leaving a clear sky. I found many familiar stars, the Summer Triangle, Mars and Antares, the Water Jar, and Pegasus rising in the east. However, the nearly full moon added its light pollution to the streetlamps and house lights, hiding the dimmer stars. It was a good time to look at the moon, however, and become acquainted with its major features as seen in binoculars. The usual star atlases have moon maps as seen in an inverting telescope, and these are quite hard to use to decipher the binocular view. There is a sketch of the principal features of the moon in the page on Lunar Mansions that may help to point out the landmarks.

It is hard to see actual craters with binoculars, but differences of brightness are very apparent. The dark area in the northeast is Oceanus Procellarum, the Ocean of Storms, with the bright crater Copernicus visible in the middle. To the southwest is the bright area of lunar highlands and mountains. Both areas are dark rock, but the maria are darker, like black volcanic ash. They look bright because of the darkness of the night. Identifying these two areas is a first step.

To the west is the line of three maria that merge into each other: Serenitatis, Tranquillitatis and Fecunditatis, flanked on the west by the circular Crisium, sea of crises, and on the east by the circular Mare Nectaris, Sea of Nectar. These are easily recognized in binoculars. Between Serentitatis and Copernicus are the Apennines, not recognizable as mountains, but as a bright linear object. Somewhere between the east end of the Apennines and Copernicus is the crater Eratosthenes, which I am not sure I recognized.

The large Oceanus Procellarum extends to Mare Imbrium, Sea of Thundershowers, on the north, with the crater Plato on its northern shore (I am not sure I saw Plato, either). On the other end of Procellarum is Mare Nubium, Sea of Clouds. Beyond it, in the bright southern area, can be seen the even brighter crater Tycho with its rays of debris, visible in binoculars.

The craters become craters and the mountains, mountains in a small telescope, and the shadows are spectacular. Binoculars reveal only brightness variations, but this is enough to become familiar with lunar geography, and the ragged terminator at other phases than full is also worth seeing.

Forest Fires

The forest fires of summer 2002 produced at least two interesting effects observable in Denver. One fire was close to Denver (the Hayman fire) and several more were 50 to 100 miles to the west. These fires produced smoke plumes that were easily detected in Denver when the wind was westerly. One effect of a large fire is the production of pyrocumulus clouds. The media were not aware of this name, but examples could be seen from Denver, and were exhibited on the TV news. Lightning was even reported from them, which could light new fires as the storms drifted east. The other effect was optical. Seen through a smoke plume, the sun was orange, and could be looked at with the naked eye without discomfort. The color of the daylight was an odd, characteristic color I have only seen when it was produced by smoke. The moon displayed a brown halo about 20' wide around it, showing small-angle scattering. The size of the scattering particles must be about a wavelength in diameter (0.5 μm), to make the scattering wavelength dependent, but not the λ-4 dependence of Rayleigh scattering by particles much smaller than a wavelength.


There is a fascinating website on contrails at Contrails.

Return to Astro Index

Composed by J. B. Calvert
Created 1 November 2000
Last revised 13 April 2003