What is this phenomena called of how the snow rolls over the tops of the trees, and how does it form? I have some suspicions based on the weather that formed them (wind and powder snow), but it'd be great to learn more about the specifics.
Here's a guess: dig into the bend or droop of one of those. I think you'll find the stem of an evergreen in there. Clear off enough snow, and the stem will straighten up... no more droop.
In other words, the snow fell so thickly and heavily (i.e., it was sticky, and didn't slide off the tree) that the tip of the tree bent over. Once that process begins, it tends to be self-reinforcing: a bent stem (closer to horizontal) offers more surface area for snow to fall on. (That effect diminishes once the stem bends all the way toward upside-down, but it's a one-way process: as long as the snow doesn't fall off, the only way the stem can bend is down.) The snow was sticky enough that it stayed in place even when it reached an upside-down position.
Details still to be worked out... why the bend seems to happen at a consistent spot along the height of the tree (though only in some trees); role of wind (direction of bend? Additional loading once in the upside-down position?); how much of the droop contains stem or was loaded when the tree was vertical, and how much is added by wind on the underside.
That's a good guess, and also what I thought, however the tree isn't bent under it... In fact these formations were also on the tops of solid, chopped-off treetrunks, too -- and also on the tops of a fencepost. These didn't have flexible tops, or anything above a flat surface for that matter.
This might be more tricky to understand than at first thought...
I wonder what the flow lines over a column of snow look like? It seems like snow is more likely to accumulate in sheltered places, so perhaps the snow is caught in vortices behind the initial snow pillar and sticks to the back side of the pillar more than the front side since the wind on the front side is strong enough to blast it off, whereas the turbulence on the back side is just enough to bring it into contact? then once the curve starts, one could imagine the same preferential loading on the bottom vs. the top since the wind would be too strong over the top. Not a very satisfying explanation. It reminds me of a formation I saw in the andes on top of a ridge - it was kind of a hollowed out sphere of snow. You could see swirly lines in it, which I interpreted to be related to the flow pattern of the wind when it formed.