Patterns in ice

[simon.snake]

During a recent cold spell (January), a pattern emerged on top of our conservatory roof after some heavy frost.



This looks very fractal like and I wonder how to recreate it in any fractal program available?

[Tglad]

Beautiful.
My guess is that ice crystals grow more easily from the edge of an already existing ice crystal than just starting on a random part of the roof.
So, it looks like the new crystals grow roughly in front of the previous ice crystal, but with a slight angle, and occasionally the conditions are right that two separate routes develop and the resulting tree branches.
Possibly the variation in straightness comes from differences in the temperature... the temperature might have dropped quickly and then levelled out as the last parts of the pattern were forming.

I doubt a fractal program would reproduce this pattern, but maybe someone has made an ice formation simulation somewhere on the web.

[fractower]

It looks a lot like diffusion limited aggregate. One possibility is that water is initially slightly salty (are you near an ocean). As ice forms it drives the salt into the surrounding water which lowers its freezing point. This will cause linear features (points grow faster) that tend to repel each other. I will try to code this up and see what happens. In the mean time could you go out and lick the ice to see if it is salty... I double dog dare you!

[simon.snake]

Afraid to say I am as far from the coast (Birmingham, UK) as you can get, and this was in January and the cold spell only lasted about a week.

I was browsing my pictures and found it, then thought of this forum and uploaded the picture and posted the message.

Will be waiting with anticipation to see what you can come up with.

Simon

[cKleinhuis]

this is exceptional beautiful!

[KRAFTWERK]

This is really cool, in many ways! 
Lovely image!

[DarkBeam]

See this site.

http://psoup.math.wisc.edu/Snowfakes.htm

  Wonderful pictures, but not easy to understand methods

[simon.snake]

Strange thing is that unlike the snowflake's quite geometric shape, this appears to be flowing and more irregular.

Was glad I took the photo when I did as a few days later it had all disappeared.

Simon

[DarkBeam]

Because here there is an "additional parameter" (gravity, temperature , whoever knows) that is = 0 in the snowflake equation.

[DarkBeam]

It maybe a mix between normal crystals and the good old fern

http://en.wikipedia.org/wiki/Barnsley_fern

[taurus]

It looks a lot like diffusion limited aggregate.

i often saw icy glass panes with the look of diffusion limited aggregate fractals, but this is different. it looks like the freezing water was still creeping for some time before congeal entirely.
a really nice structure. great find simon.snake! 

[simon.snake]

The roof of the conservatory is glass, and slopes away from the house at maybe 20 degrees downwards, and is only single glazed.

The picture was taken from the window above the roof so the top of the picture is lower than the bottom of the picture.

Seems like some of the fern like structures are going uphill???

Simon

[simon.snake]

The original picture is much larger than what is shown on the post, so open in another window and zoom in if you like.

Simon

[tit_toinou]

I have taken the same kind of photos last winter.
This one is awesome  .

[Mrz00m]

water is one of the few materials that expands when colder... it shrinks until it approaches zero and then when it freezes, the random mobile molecular structure becomes patterned into straight lines, with a honeycomb crystal form, that's why it tends to brach off at 30 degrees angles, esp. in snowflakes.