Water plane looks different inside/outside the octree

The water plane looks different when it is outside of the octree compared to inside the octree but no chunks. Orignally reported in Discord: https://discord.com/channels/541221265512464394/1156386708145848451/1156386708145848451


@ThatRedox already investigated this issue a bit:
> looks like rays that intersect with the water plane will continously enter and exit the water octree
> Also rd<=4 will not have the weird artifacts in the octree
>
> ok so it looks like its going something like:
> 1. Initial ray going down
> 1. Hits water plane -> air-water refraction
> 2. Hits air in water octree -> water-air reflection, now going up
> 3. Hits water plane -> air-water refraction
> 4. Hits air in water octree -> water-air refraction
> 5. Hits sky

Reproduce the water-plane comparison described in the issue, both outside the octree and inside it with no chunks, then trace the ray path through the water plane and water octree. Use the reported repeated air-water and water-air transitions as the starting hypothesis; the issue is done when both contexts produce the same appearance without the reported artifacts.

Written by the indexing model from the issue text.