Noctilucent clouds (NLC) are the highest clouds in the Earth's atmosphere, laying at the average height of about 83 km in the mesopause. They are typically observed only within 50-65 degree latitude region in both hemispheres, and only during local summer. One can typically see 10-20 noctilucent-cloud displays in a year; thus they are rare.
Noctilucent clouds look a lot like high cirrus clouds, but they are a lot thinner optically, and bluish or silver in colour. Actually, they are optically so thin that they can only be seen during twilight hours, when the sun no longer illuminates the ground or the thick lower atmosphere, but still shines upon them. These conditions are met only when the sun is about 6-12 degrees below the horizon; less than that means that the sky is too bright, more than that and the noctilucent clouds are in the Earth's shadow as well. Like cirrus clouds, noctilucent clouds are generally thought to be (at least mainly) composed of water ice.
There are actually two rivalling theories for the noctilucent cloud formation and composition, namely the ice theory and the dust theory. The latter proposes that noctilucent clouds are formed from cosmic or volcanic dust, but this theory has several shortcomings and is generally thought incorrect. The ice theory, on the other hand, is consistent with observations and can explain the observed climatology (fairly constant and restrained altitude, appearance only at high latitudes, and only during local summer) of noctilucent clouds. Thus, only the ice theory is explained in detail here.
According to the ice theory, noctilucent clouds are composed of water ice. The cloud particles are formed either by heterogeneous nucleation on small dust particles, or by a so-called ion-induced nucleation on certain ions in the polar mesopause.
Nucleation can only take place if there is super saturation relative to, at least, ice; this demands a very low temperature, as the whole middle atmosphere (stratosphere + mesosphere) is very dry. Also, there must be a continuous source for water vapour, because the hard UV radiation breaks down water molecules in the mesosphere. In addition, the formation and the growth of cloud particles uses up the water supply if it is not continuously replenished. These are not conditions easily met.
One would think that it must be very cold up there as temperature decreases upwards, but this is only true for the troposphere, the lowest part of our atmosphere. Above the troposphere, the temperature begins to rise again, mainly due to absorption of UV radiation by ozone. Still higher in our atmosphere, even more energetic (i.e. shorter wavelength) radiation is absorbed; thus the upmost part of our atmosphere is actually the hottest (it must be pointed out that air is very thin at these heights, so the very concept of heat is a bit vague there). So, in the middle of this solar-powered furnace, how could there be a very cold place for the ice clouds to form?
The answer to this question is surprising - the attenuation of vertically propagating gravity waves (or buoyance waves; these are not the waves of gravity field the astronomers are trying to discover). These waves are mainly caused by jet streams, thunder clouds, and mountains. Vertically propagating waves transfer a lot of momentum from the lower atmosphere upwards. The transfer of momentum takes place when waves attenuate; this can happen by radiative cooling or convective overturning (if waves are unstable). It turns out that gravity waves are the only waves that can propagate upwards to the summer mesosphere, and not even all kinds of gravity waves but only those that have strong east-moving phase velocity. As they propagate higher, their amplitude increases, until they eventually become statically unstable and break up, transferring their eastward momentum to the mean flow. In the atmosphere, winds and temperature gradients are interconnected; thus the introduction of momentum also changes the local temperature field. This mechanism cools the summer polar mesopause so strongly that the region is actually the coldest place in the Earth's atmosphere, with temperatures typically around 130 K; about 60 K lower than in the winter polar mesopause where the solar heating is absent!
This extraordinary temperature distribution can be explained by strong vertical motions, driven by the attenuation of vertically propagating gravity waves. The input of momentum in the summer polar mesosphere decreases the westward flow that should occur if the mesosphere were in a radiative balance and turns it eastward. Accordingly, the coriolis effect turns the flow toward the equator, forcing a strong upward motion and associated adiabatic cooling in the summer hemisphere polar mesosphere, and compensating downward motion and adiabatic heating in the winter-hemisphere polar mesosphere. This circulation depends also on stratospheric conditions, as they largely determine the kind of vertically propagating waves that reach the mesosphere. Thus, changes in stratospheric winds, for example due to changes in uv-absorption, can significantly change the conditions in the mesosphere also. This middle-atmosphere circulation system is quite complex, depending on radiation, chemistry and fluid dynamics, and full of feedback mechanisms that make the whole system strongly nonlinear.
In addition to providing the low temperature, vertically propagating gravity waves also provide the continuous supply of water vapour from the strarosphere with the forced upward motion. Thus, the gravity waves are essential for the noctilucent cloud formation. They also give the clouds their usual shapes.
There is also another water vapour source with approximately equal importance to the transport, namely the photodissociation of methane in the mesosphere. This process produces water vapour for noctilucent cloud formation in situ, i.e., it doesn't have to be transported to the place of formation. This is an advantage, as water vapour is also photodissociated above 65 km altitude. In the mesopause, the average lifetime of a water molecule is only 3-10 days.
The above mechanism also explains why noctilucent clouds can only be seen during summer and only at high latitudes. The conditions are suitable for ice particle formation and growth only during summer months, both for the temperature and the humidity, as only the summer stratosphere allows the right kind of vertical waves to propagate to the mesosphere that provide the necessary influx of eastward momentum. During winter the zonal winds in the stratosphere have the opposite direction, so the eastward propagating waves cannot propagate to the mesosphere. The upward motions associated with the influx of eastward momentum near the mesopause only occur at high latitudes of summer hemisphere, confining the formation of clouds there. Near the pole, the latitude region of ground observations is limited by the sun-elevation angle being too high during the summer months, but satellites have frequently observed mesospheric clouds covering the whole polar region (these clouds are called polar mesospheric clouds (PMC); NLC's are considered to be PMC's ragged edges). The sufficient conditions (low temperature and high humidity) are only reached in a shallow layer in the mesopause, so the vertical extent of the clouds is always small. The altitude of the mesopause varies little, and so does the altitude of NLC's. Thus, the ice theory is very successful in explaining the noctilucent cloud observations.
Noctilucent cloud particles are probably very small - typically smaller than 50 nanometers in radius. There is little direct evidence of the particle size, but they appear to be Rayleigh-scatterers at visible wavelengths. The polarization state of light scattered by noctilucent clouds is similar to the polarization state of the background sky, which is difficult to explain with anything but cloud particles clearly smaller than the wavelength of radiation. The cloud particles can be either hexagonal ice, cubic ice, or even amorphous ice; all forms are possible in the temperatures and air pressures associated to their formation.
First noctilucent cloud observation is from year 1885, 15 years after first mother-of-pearl cloud observation and 2 years after explosive volcanic eruption of Mt. Krakatoa in Indonesia. It has been speculated that this first sighting was due to the eruption, ejecting a lot of water vapour and dust to the middle atmosphere. This is consistent with observations; following the event there were numerous and brilliant NLC displays, but then the clouds practically dissappeared for decades. Since the 20's the observations seem to have become more and more frequent, however.
It is very likely that the whole phenomenon should be credited for industrial activity. The atmospheric methane concentration has more than doubled since the pre-industrial era. The increased amount of methane and other greenhouse gases have increased the temperature in the lowest parts of the atmosphere, but decreased it elsewhere. It is thus quite possible that noctilucent clouds (and mother-of-pearl clouds) are first visible signs of our impact on the Earth's climate. Comparison of the NLC observations with the solar or aurora activity suggests that these temperature-altering phenomena clearly affect the number of NLC sightings, further confirming the ice theory.
If the situation is this, it is easy to predict that we should expect more frequent and more magnificent NLC displays in the near future. Also, the length of the NLC season should increase somewhat, and their region of existence extend closer to the equator. Unfortunately, this is not the whole story. The effect of ozone depletion on the middle-atmosphere dynamics is still uncertain, and potentially a lot more important than the amount of methane and other greenhouse gases in our atmosphere. Major ozone depletion could even lead to reorganization of the whole middle-atmospheric circulation system, among other things. The strong nonlinearity of the middle-atmosphere circulation makes predictions also difficult.
Mesopause is one of the most difficult locations in Earth's atmosphere to observe. Satellites, lidars, and radars, for example, can be used to do remote sensing measurements, but In situ measurements can only be carried out using expensive rockets. Remote sensing methods cannot provide direct measurements, and a rocket provide data with very limited spatial or temporal resolution (they penetrate the thin NLC layer with high speed in a few seconds, and only in two highly localized points, once on their way up and once on their way down).
Thus, it is not surprising that human observers are valued greatly in NLC research. A network of observers can provide valuable climatological data from noctilucent clouds, for example. Most valuable observations are those made under unusual circumstances, for example, outside the NLC season (early spring, late autumn, or even winter) or at unusually low latitudes (equatorward of 50 degree latitude). Photographs of such occurrences are valuable, as well as other means of making sure it is genuine phenomenon (such us using a polarization filter or checking the sun elevation angle).