Explanation from Barrie Jones, Open University

From: Barrie Jones


Shadow bands are well documented in a qualitative way, and there is some quantitative data, a good deal of it my own. At several eclipses I have recorded the light intensity variations with an array of photodiodes. Analyses of these variations lend support to the 'scintillation' theory. The general idea of this theory dates back many decades, but it was worked out in some detail by J Codona in the mid 1980s.

The bands result from the illumination of the atmosphere by the thin solar crescent a minute or so before and after totality. The atmosphere at any instant is an optical medium in which the density varies slightly from place to place. This means that the light from a distant point can reach a particular place on the ground by a variety of paths, each one refracted differently through the atmosphere. Thus in some places the light waves reinforce and the light level is enhanced, whilst in other places the waves tend to cancel each other out and the light level is reduced. When the effects of all the points that constitute the solar crescent are added the outcome is a ragged banded pattern of light and shade - shadow bands. The bands' contrast is low, as is the general light level near to totality.

The density variations in the atmosphere are not fixed - they evolve of themselves and they are carried along by the winds. Therefore, the pattern moves across the landscape at a speed that depends on the wind speed. If the wind is fast the eye might not be able to pick out the pattern; if the wind is very slow then the pattern will not move much, just evolve 'on the spot,' and again it is difficult to pick out. In other words, there is an optimum flicker rate for the eye. The wind direction also matters. If the wind blows along the bands (rather than perpendicular to them) it is not very different from having no wind at all. For bands spaced by 0.1 metre or so, a wind component of a few metres per second in the km or so above the ground in a direction perpendicular to the bands is likely to make them visible.

The contrast of the bands is said to be better if a filter is used, but then the light levels are lowered overall, so filtering is a delicate, uncertain balance.

Elevation makes little difference unless you go to over a km or so, in which case the amount of atmosphere above you to produce the bands has been significantly reduced, and the bands are fainter. Haze can also reduce their visibility.

The low contrast, low light level, and motion, all combine to make it very difficult to obtain sharp photographs of the bands. There are some impressive videos, and a handful of still images. However, for scientific work a few really sharp frames are needed, so that the two dimensional pattern of the bands can be examined at a resolution of a few millimetres on a white diffuse screen a couple of metres from the camera.

For a review of my work see Journal of Atmospheric and Solar-Terrestrial Physics, volume 61, pages 965-974 (1999).

Barrie W Jones