This question is inspired by the following comment on 
https://space.stackexchange.com/questions/5023/would-colonising-antarctica-be-a-good-test-for-colonising-mars#comment13675_5023
:

> This isn't a direct answer, but I think colonising underwater would be a better practice for colonizing Mars. Some problems are reversed in direction (you're trying to keep water out rather than air in), but the challenges are similar: you can't breathe outside and have to take precautions going in and out, reduced sunlight of similar length, isolation both physical and communication-wise (assuming there isn't an undersea cable), and unusual hazards.

Suppose we had a manned capsule that could dive into the depths of the ocean at a controllable rate. What acceleration would it need to maintain to simulate Martian gravity within the capsule?

If the capsule were diving to the deepest point in the ocean (roughly 7 miles or 11 km), how long would the inhabitants have in the simulated Martian gravity?

Suppose the capsule could also ascend at a controllable rate. If it turned around at the deepest point and ascended, how much additional time would this give in simulated Martian gravity before reaching the surface? I would think it would double the time, but I'm not sure.