Have you considered the effect of tides/water movement?

One method in use is to generate a magnetic field then measure the voltage induced into a pair of sensors by the water flowing through the field (it being the conductor). Obviously using the sub in fresh water will produce some problems.