Speed of sound in water seems to be ~1500M/sec, so a 10cm seperation (just to make the math easier) would be a ~7uS difference, if the source is aligned with the sensor axis. In order to get a 30 degree resolution, you'd need 7uS * (1-cos(30deg)) = ~900 nsec.
For a 200KHz signal (at the 10cm spacing), this is a phase difference of 60 degrees (and 500 degrees in the best case). Tweak the spacing between the receiver elements and the pulse frequency so that the phase difference never exceeds a wave.
If the signal is strong enough, and you use the InputCaptureModule (ICP) with some simple logic - an XOR gate fed with a pair of mux's so you can select which channels you want to monitor, it should work.
A resonant tank circuit on each channel - tuned to the pulse frequency (200KHz in this case) would probably be a good idea as well.