OK sure. What do I know...I agree that this is an eventual failure mode, and thank you for pointing this out. To turn your concern into a design spec, I'll restate it this way: we need to figure out the depth of water such that the cabin displacement is large enough to balance the mass of the vehicle (neutral buoyancy).
Here's the math: I estimated the interior volume of the FJ at 100 cu feet (it's 70 CF cargo space with the seats down, so I think that's reasonable; if anyone knows the actual number please tell me). For a 5200 lb FJ and subtracting out buoyancy of the (my) tires, you have net mass of 2000kg, which requires 2000 liters of cabin volume under the waterline for neutral buoyancy. That's 70 CF, so at the point that the FJ starts to float is when all but 30 CF is submerged.
The FJ is roughly 8' x 4' around the top, so a foot down from the roofline is about the spot. Note that this is true no matter what the actual height of the vehicle is.
Note further that this is a little higher than the empirical data point we have from Bluecuss about mid-windshield being where things got sketchy. You would expect things to start getting increasingly sketchy as you got within a few hundred pounds of neutral buoyancy; the hydrodynamic forces from any movement of water would start to push you around more and more.
I wonder if there are FJs out there with enough extra gear that could submerge completely and not float. It would take another thousand kilos.