Charlie J,
First off, I hate putting electronics in water. Especially sea water.
That said, how accurate does the speed data need to be?
Since the pump no doubt has a data sheet describing the flow rate at a prescribed DC current rate, you can monitor the current (Amps) the pump is using to extrapolate, quite accurately with the appropriate algorithm, the flow volume.
If the power source for the pump is external (above water), then this could be done without submerging the sensor. It can be done with a extremely simple circuit (a simple resistor) that senses the current flow to the motor which could then be supplied, through an A/D converter, to LabVIEW. I can provide a schematic, if you like.
Even if the power source is at the pump, it could still be done: just need to provide a connection to the current sensor.
Had great success with this concept monitoring the current usage (and, therefore, general effectiveness and flow rates) of pumps providing aeration and water exchange for a commercial shrimp aquaculture operation.
AC was, of course, the power source (and we used a BUNCH), and the sensors were current transformers, but the concept is the same.
It also had the secondary advantage of providing really basic information like, "is the thing running at all" or, "has the impeller gotten loose", for alarm purposes.