I think a few general rules like this are the best you can do, especially without trying to map out the entire sea floor and run a resource-intensive simulation. We can certainly say that certain proposals would definitely not be at all likely, so the rules can keep from creating something implausible, but I think there's still going to be a lot of wiggle room to assume things flow in ways that are productive of the situations desired by the author.
The formation of deep water is governed by density, which is a balance between salinity and temperature, both of which are governed by wind-driven surface evaporation. So...
1) Fast wind and freezing temperatures both cause water to become colder and saltier, and therefore to sink; hot rainy climates make it warmer and fresher and therefore less dense. (Hot dry weather and cold rain both have less effect, obviously)
2) The water that sinks to the bottom flows downhill, and flows away from the points where it sinks from the surface; water is "produced" at the rainy tropical latitudes and flows toward the points where it sinks.
I don't know what you can do with this -- I have to go do other science but I'll be back later. :-P