Turbulent Sirius ~ Defocused images of Sirius by Peter RosÚn showing fast moving and changing colours from atmospheric refraction. He was attempting to image Sirius B (see images below) but when Sirius had sunk to 10° the session was curtailed by atmospheric turbulence.

"I decided to study the increasing turbulence instead.

I switched camera to my color CCD, defocused the image by quite a large amount and took a video sequence of about 100 frames at a rate of 3.75 frames/s. They show a disc with rapidly changing colors on every frame. The telescope spider vanes, the secondary and the holders of the primary mirror are seen as a black shadows."

4x Televue Powermate on a CT-10, 250mm Newtonian from OrionOptics.

All images ©Peter Rosén
Starlight caustic sheets from a wavy warm-cold air boundary extend downwards to touch the surface. Illustrative only - not to scale.
About - Submit Optics Picture of the Day Galleries Previous Next Today Subscribe to Features on RSS Feed


"At first I pushed the telescope during the exposure so it would start wobbling leaving a Lissajou-like colourful path that clearly shows the fast changing colours and luminance of scintillating Sirius. 

This is nothing new as it has been done several times before but I took this image and superimposed it at the correct scale on one showing the star field. Canon Eos5D MkII and a 4x Televue Powermate on my CT-10, 250mm Newton from OrionOptics."

The changing colours result from refraction across moving layers or pockets of air at different temperatures.  They disperse the starlight into its spectral components and deflect it in different directions.    The effect increases as a star gets lower because its light takes an ever more slanting path across the air layers. They are the bane of ground based astronomy.

Some twinkling and changing colours, but not all, come from caustic surfaces sweeping and rippling across the eye as they move and intersect the ground level in an ever changing pattern. The varying network of bright lines on the bottom of a sunlit swimming pool or shallow sea is closely analogous.

The computed ray tracing at right shows an example of how light passing from a less dense medium (warmer air for example) via a wavy interface into a denser one (cooler air) is deflected. Stronger refraction than that of air was used to fit the rays on the screen! Some distance below the interface there are pairs of surfaces where two rays crossing each other cluster closely together. These are caustic sheets and they are regions of intense light.  Sheets of different colour are slightly separated.  As the sheets sweep across the eye the star momentarily gets very bright – it twinkles and in extremes flashes colours.

"In this step I stacked 100 colorful discs to see what would be the average color of the sequence and got a result that is almost completely achromatic with the exception of some color [diffraction] fringes.

Stacking the 100 images was like inverting the prism that split light into a colorful spectrum. 

I read somewhere a long time ago that the medieval artists who composed the stained glass in churches and cathedrals, had a good knowledge of how to balance the amount of each color so the resulting light would not be tinted. This was a long time before Newton."


"I have a long running relationship with Sirius as I have been trying to photograph the elusive Sirius B for at least 3 years, not very successfully until now.

At a maximum altitude of less than 14° at transit time, I think that it has never been done by an amateur in Sweden and I am told that it is probably impossible. The separation of 10 arc seconds between the two stars is no problem but the difference in brightness and the low-altitude turbulence makes it an extremely difficult task.

In my latest attempt [Image A at right] there is a bulge on Sirius at the correct angle and distance on the left side but visible only on 2 frames. So I will wait until I can confirm it before claiming it to be Sirius B.

[The larger blue image B] is work on very different images taken in February last year that might also show Sirius B. But again, it might as well be an optical artifact.  I used a CCD with my 250 mm Newton telescope and added layer upon layer, masking the brightest part so as not to overexpose the core.

I have also used a frequency modulation technique to enhance the contrast in every part of the image, thus making for a positive-negative effect. The spikes are from the telescope spider vanes and from 2 different sessions with the telescope rotated 45° totaling 8 spikes. The blue color was added purely for aesthetic reasons as the images were caught with a monochrome CCD.

Although I am not sure that the tiny white dot directly to the left of Sirius A [in red circle] is really Sirius B, it illustrates nonetheless the incredible brightness of the Dog Star and the difficulty in separating the two, especially at a maximum altitude of only 14° above the horizon."