You also have to remember sampling - an Ohio is approx 170m x13m x 1m
To get an "anomaly " you're going to need 2 or three "anomalous values minimum - so that's approx one every 50m if you are lucky enough to be running parallel with the boat and every 3m across track
At 180 mph (80 mps) you have to get a value every 0.5 sec or 0.04 secs depending on relative heading. Worse at jet speeds of course.
Then you have to locate that anomaly in a bucket load of data (after several hours of ongoing data collection) with all the issues of "noise" mentioned above in real time.
I42 will correctly point out that the Ohio anomaly would be bigger than the actual vessel - but that just spreads it out across a little bit more ocean , with the edges almost impossible to define. Every grav survey I've seen is pretty good at defining things down to 3-10 kms resolution - after that you're kidding yourself.