Sprites are a fleeting, ethereal and a relatively unknown aspect of lightning storms.
Since the 1960s, and probably before then, pilots have been seeing but seldom reporting what have become known as sprites and elves above the clouds. Sprites are electrically-charged lightning funnels which shoot up from the top of a cloud as much as 60 miles into the atmosphere. These charges are vivid red and usually occur in clusters of three or more but are only visible for nanoseconds. They are sometimes preceded by lower altitude red flashes known as elves, and can have striking blue tendrils which are easily mistaken for blue jets. While they are a similar visual phenomenon, blue jets are less powerful than the sprites and travel neither as quickly nor as far.
Because ‘everyone knows’ lightning goes to ground, pilots were naturally reluctant to report this phenomenon in case they found themselves grounded for hallucinating. As a result, serious research was delayed until the last 15 years or so.
While sprites are more common during positively-charged lightning storms, this is not due to any preference on the part of the sprite, but rather due to the greater internal energy of a positively charged storm. It was not until 1999 that the first sprites of a negatively-charged storm were recorded.
During a powerful storm it is possible to see red sprites, elves and blue jets, but the exact atmospheric conditions which create such a show are uncertain.
As sprites are relatively new to the science world there is still a lot more to learn about them.
It is only with the advent of high speed photography that the existence of these light shows could be confirmed, and even with that they were first photographed by accident in 1989. Amazingly, there have since been more than 10,000 confirmed sightings. They are also known to create a very low-frequency thunder which was only recently captured with the use of specialist listening equipment.
As their energy is spread more thinly than the traditional thunderbolt due to the cone like dispersal from cloud to atmosphere, they are thought to be relatively weak. Sprites are cold plasma phenomena that lack the hot channel temperatures of tropospheric lightning, so they are more akin to fluorescent tube discharges than to lightning discharges.
The effects of sprites are currently being investigated by various agencies including NASA who seriously addressed them as a possible cause for the tragic loss of the space shuttle Columbia, which was, incidentally, on a mission to record data about the very same sprite phenomenon.
The link below shows footage from the ISS and shows a red sprite over East Asia at around 0:06.
Sprites are sometimes preceded, by about 1 millisecond, by a sprite halo, a pancake-shaped region of weak, transient optical emissions approximately 50 kilometres (31 mi) across and 10 kilometres (6.2 mi) thick. The halo is centred at about 70 kilometres (43 mi) altitude above the initiating lightning strike. These halos are thought to be produced by the same physical process that produces sprites, but for which the ionization is too weak to cross the threshold required for streamer formation.
Recent research carried out at the University of Houston in 2002 indicates that some normal (negative) lightning discharges produce a sprite halo, and that every lightning bolt between cloud and ground attempts to produce a sprite or a sprite halo.