The title says it all...

enter image description here

I assume it requires very slow freezing sheltered from the wind. Undercooled water, maybe?

  • A Reddit thread talks about Delaunay triangulation, but that does not explain how it forms.
  • Many sites claim it is in Switzerland without further info.
  • A comment in another Reddit thread states It happens when there is a significant change in gravitational force during the freezing process.
    Also, there was a lot of water vapor hovering just above the surface during a rapid freezing process
    , but I doubt that gravitational effect.

BTW While researching this I found this beautiful page 18 Beautiful Frozen Lakes, Oceans And Ponds That Resemble Fine Art - several of the pictures there would merit a separate question ;-)

It has a similar picture claiming to be from a frozen pond In Southern Oregon, USA*:

enter image description here

* But a Google image search finds claims of Switzerland, again.

  • $\begingroup$ I have seen similar features on a much smaller scale on frozen ponds. If I recall correctly it seemed to occur near or above bubbles of trapped air but don't hold me to that. But perhaps it has something to do with the ice forming while the water below is contracting in volume as it cools below 4C. $\endgroup$
    – haresfur
    Oct 12, 2015 at 2:50
  • $\begingroup$ I believe the ice cover is real as there's a handful of similar pictures from different lakes/ponds. Ice can form in neat geometric shapes with sharp angles with high humidity where it gets layered neatly. Snowflakes are geometric based on that principal, but I can't give and couldn't find a good explanation. I agree with you on low wind and calm water as possible to likely factors. $\endgroup$
    – userLTK
    Oct 12, 2015 at 7:46

2 Answers 2


Yes, it is real. Whoever took the photo, congratulations on a very fine image. I have never seen this texture on such a scale, but something similar can be achieved in the laboratory by creating a bubble membrane of pure super-cooled water and blowing a few dust nuclei onto the surface. A comparable geometry appears very rapidly. The size of the ice crystals are difficult to judge without any feature to give scale, but they seem to be exceptionally large trigonal terminations. We have all seen snowflakes of hexagonal symmetry, but under metastasble conditions with rapid crystal growth, stacking disorders tend to occur in the crystal lattice, yielding a trigonal rather than hexagonal symmetry. See, for example, The ice phases of water. This is an unusual polymorph of ice, and I cannot see how it could be achieved except through sudden seeding of very pure water. Perhaps stratified freshwater lakes, in still air, on a very cold night, with a surface of previously melted ice (i.e. pure water), and subject to a sudden windy gust of nucleation particles would provide the right conditions?

  • $\begingroup$ Regarding trigonal 3-fold symmetry, check out the most recent issue of BAMS (Sep 2015). Murray B.J. et al, 2015: Trigonal ice crystals in earths atmosphere. Bull. Amer. Meteor. Soc., 96, 1519-1531. $\endgroup$
    – casey
    Oct 16, 2015 at 17:43
  • $\begingroup$ Would multi-vertice (>3) centers indicate nucleation sites? And the apparent vertical relief, is that real, or an optical illusion? $\endgroup$
    – Eubie Drew
    Oct 24, 2015 at 20:41

This picture was taken in what seem to be a small pond, of very calm water. It seem that in ideal conditions, implying but not limited to :

  • absence of winds
  • clean water eg: no nuclei to provide an anchor for crystal formation
  • low temperature gradient between air and watrer (air almost near 0 °C and very slowly falling)

ice grain can grow quite large.

This paper discusses the matter of ice classification for lakes and river. Fig. 1 display a flow chart on possible paths for ice formation - I would interpret that the ice formation followed the left path, from fresh water, to calm water toward plate-ice, followed by undisturbed growth to ice sheet, and ice cover.

To cite how the primary ice would form on the pond (Michel and Ramseier, 1971) :

On a calm surface the primary ice is an ice skim which grows horizontally in the supercooled layer and is a few tenths of a millimeter thick.

For a calm surface and a small temperature gradient, it is possible to see forming very large to giant (up to meter sized) grains in ice. Later in the discussion :

Many of the platelets or discoids develop into dendrites, and the interaction of the dendrites causes the crystal boundaries to be very angular and elongated in shape.

Michel, B. & Ramseier, R. O. Classification of river and lake ice Canadian Geotechnical Journal, 1971, 8, 36-45

The following is a reference on the matter in my opinion:

Shumskii, Petr Aleksandrovich. Principles of structural glaciology: the petrography of fresh-water ice as a method of glaciological investigation. Dover Publications, 1964.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.