Unless you're already seeing performance issues, I'd currently not worry about it and build the rest of the game first. There are enough challenges ahead, getting hung up on an optimization that isn't currently causing an issue, might just slow you down too much.
There isn't necessarily just one possibility.
The simplest approach would probably be to keep a VA per chunk.
But you could also dynamically stitch together multiple chunks into one VA. It's a question of memory usage vs performance.