How can such a mushroom ice shape grow from a water fountain? Recently I stumbled upon a funny picture of a frozen water fountain. The frozen fountain resembles a mushroom:

Source: https://izismile.com/2021/02/19/this_is_extremely_rare_50_pics.html
The longer I look at this picture, the less I understand how such a shape can even start growing.
While there is still flowing liquid water, the only place where this icy shape can start to grow is the metal border at the bottom. But then, how is it growing there? Is the ice forming a border first, growing upwards and finally freezing to the middle?
And how and when did the "mushroom head" grow? I assume the water was in permanent flow.
It just hit me that the "mushroom head" was probably not shaped by the water stream becoming a membrane, but by the metal shape below. The relief of the "head" seems to match the relief of the metal below. But then again, how was the "mushroom head" lifted once it was frozen? If still water in the tube was freezing and expanding, was that really enough to lift so high?
I can't explain it. Therefore my question:
Can somebody explain how this ice shape has been formed, out of flowing water, metal and cold?
 A: Just a guess:

*

*The head is formed by freezing water that's in contact with the metal and covered its surface.  The shape of the head fits the shape of the metal.

*Presumably, water / ice was still pushed out of the opening as it was freezing inside the metal:  ice has a lower density than water, thus the same amount of ice requires more volume.  This pushed-out water / ice then formed the trunk which lifted up the head.  It's likely some reservoir in the metal part, which is the end of a (longer) pipe.

Also note that ice melts under pressure.  This might also play a role as it makes it possible to push the ice out of the opening (which temporarily mety).  Or it was just expanding (cooling) water in the pipe, that only froze as it was slowly pushed out of the opening.
And it might be the case that the valve further down the pipe might not be 100% tight and releases some droplets each minute or so:  First that water just wettens the metal, and when freezing starts the head is formed.  Aterwards the additional water forms the trunk and pushes the head.
Whatever mechanism is at work, IMO it's that 2-stage process at work.
