Actually, the mixing properties are incredibly important. Even under the ocean much oil isn't in big cavities, it's in porous rock. If this is the case with fre

Actually, the mixing properties are incredibly important. Even under the ocean much oil isn't in big cavities, it's in porous rock. If this is the case with fresh water, the structure containing the fresh water might actually sufficiently support the salt water on top of it as we reduce the internal pressure when extracting fresh water. Even if not, we can drill and extract fresh water at one end and pump in salt water at the other, until the water we get out becomes too saline. In porous rock the mix rate would be quite low; kind of like pumping salt water into a sponge at one end and extracting the fresh water at the other end (for a while). This is quite a common practice in the oil industry. Finally, salt water is more dense than fresh water. If these reservoirs underwater actually are big cavities; we can just pump salt water in the bottom and pump fresh water out the top. NB: When I say "we can just", I mean, take a wild best guess about where the bottom and top of the well is based on crazy amounts of uncertain data extracted on a well-by-well basis.