Thursday, 2 May 2019

Kern with subhelic arc in the UK

Berkshire (UK), 28 April 2019: Top left corner shows what's left after applying the gradient subtraction and stretching of histograms on a 38-frame average stack. Top right and bottom left, respectively, are with additional blue-minus-red subtraction and colour-channel enhancement. A HaloPoint simulation is included for reference at the bottom right corner.

While relaxing at home in the evening of a mostly cloudy day, a ray of light caught my eye and I checked the halo situation. In fragmented cirrus there was a mediocre circumzenithal arc (CZA) but not much more than that (my view towards the setting Sun is not that great). I took notice of the Sun being low in the sky so odds for Kern arc were on the rise: in the next moment I was setting up my DSLR to get some photos from my backyard.

CZA disappeared soon after I started shooting the first set of images, but it came back a few minutes later so I chose to do a re-run. Fortunately so, as the latter set turned out fruitful indeed in the post-processing.

The processed stack indicates presence of a faint Kern arc as an extension of CZA at the left. Slightly further to the left, there is a white arc that best matches the subhelic arc in the simulation. I think there is also a subtle suggestion of helic arc alongside of the subhelic, but this is less convincing. Furthermore, there are colored patches below the CZA approximately where 46° contact arcs would appear if there were Lowitz-oriented crystals in the mix. The simulation shown above is a synthesis of three distinct populations, including singly-oriented plates, singly-oriented columns, and Lowitz-oriented plate crystals.


  1. Absolutely first rate, congratulations Reima! I think both the helic and 46° contact arcs are definitely there and are VERY compelling. Indeed, with regard to the Kern, not only is there an extension of the CZA on the left hand side, I also see an almost complete circle, especially on the br processed image. You can even see where it touches the subhelic, exactly matching the simulation.

  2. Additionally, I also think the D-stretched version amplifies the very faintest details in the image extremely well and should be a regular tool in our image processing armoury. For the record, what colour space did you use in D-stretch?

  3. Thanks Alec. Sorry for making you confused, but the post includes no D-stretched version. Lower left panel is just with enhanced colour channels using Gimp.

    1. Don't worry, I'm very easily confused! Mea culpa!

  4. Yeah, pretty much full circle there. I expect the next high cloud Kern come from Eresmaa again.

  5. Its amazing the wonders of image stacking