One question is, how should one understand the GPU vram usage, when it is reported as 16.0/15.7 GB?

IIRC, using more vram than what's on the card isn't a show-stopper, just slow. The parts that don't fit on the card just lay in RAM and are swapped back on card when needed. This is fairly slow and comes with a performance penalty, which is why the numbers are shown as red on there.

as for why it's not showing full 16 GB as being the max? ... no clue, probably the card needs some vram for it's own operation/general framebuffer (or whatever is the term) for displaying eg. the os, not just the game. But I'm just guessing.