A couple of weeks ago, I got a cool picture of snow hanging off SteelyKid's playset, and posted a call for people to suggest physics-y ideas about that. I only got one response, probably because nobody really read the Internet over the holidays.
Anyway, the next time I'm likely to have the free time to write anything substantive is next Tuesday morning, so let me renew that call now, while people are trapped indoors by the POLAR VORTEX! with nothing to do but think about the physics of cold things. So, if you have physics-related ideas about the photo above (there's a view from a different angle at the link), send them to me by next Monday, and I'll write something up on Tuesday.
I suspect you are seeing the same sort of thing as the slabs of snow that are involved in avalanches. A slab of snow develops a binding of some strength, and in the case of an avalanche a weakness zone develops on the bottom. At some point the weight of the upper zone exceeds the weakened strength of the boundary and the avalanche occurs.
The picture shows a slab of snow that has frozen together a bit (likely it was a wet snow) and things got colder so it consolidated at least somewhat.
Then you notice that the cloth has a sag in the middle between the supports, that was there before the snow slid. The sag was enough to halt the slide at least for a while.
If you read about behavior of glaciers as well as avalanche control a lot of work has been done on these subjects.
I see 2 interesting questions:
1. Why snow remains whole after going off the roof?
2. Why it bends, not breaks?
I think, snow is transparent, light warms the roof, wet snow at bottom freezes at night making an icy base. After sliding a little off the roof part of this base near the edge of the roof is exposed to a warm water flow. It melts the base, but in wet snow surface tension keeps it from breaking.
Addition: when snow starts breaking the superdense packing of snowflakes in it becomes much more loose at once. There appears more room for water and it becomes a thin film on snowflakes. That converts mechanical energy to surface potential and therefore stops breaking.
I don't like the explanation, but don't see better one.
I disagree with Mike, above, about this being related to refreezing of snow. I say this because I see similar things on my own roof (particularly the part above the garage, which is not heated; I am less likely to see these above the main part of the roof, which has only an attic between it and heated living space) immediately after certain snowstorms, including the one of 2-3 January of this year (when it was unusually cold for snow to be falling at my location). Obviously, the structure in the photo is also unheated. My best guess is that it's an effect of wind interacting with the structure, something akin to drifting snow. But I don't know for sure. Maybe if somebody who reads this blog lives in the Upper Midwest/Great Plains/Canadian prairie, (s)he could say whether there is any merit to my guess.
Well, I agree with Eric :) on such wind-created formations exist. But all of them have a somewhat horizontal bend, while the thing in question is strictly vertical. And there is sign of the snow layer shift as a whole (clear part at roof top).