I've been reading ULB manuals and patents I'm still confused about the ducting of acoustics in the water. I understand that temperature and salinity levels, water surfaces, ocean bottoms, etc. can cause the sound to be refracted, or reflected or whatever; and I keep reading that water will attenuate the ULB pings at roughly 5 to 7 dB per km, depending on depth, salinity, temperature, etc.. I'm guessing that the ducting waterways we hear about don't consist of 'magic' water so the pings still get attenuated at the same 5 to 7 dB per km, just perhaps bounced in odd directions.
Is that about right ? If so, then the source has to be pretty close to the detection areas. Even if the sound is being bounced around in crazy directions, it's still getting attenuated at a known rate as it moves. I can see if the ULB fell into a parabolic shaped canyon and the canyon walls were acting as a focusing (antenna)transducer, maybe the range could be a little further but it still seems like it has to be in the right neighborhood. Maybe they haven't found the haystack yet but they very likely are on the correct farm.
Here's an old test the FAA did back in 1968 when they dumped a fuselage in the water off the florida keys in 200 feet of water to see if a fuselage encapsulated ULB (inside the pressure bulkhead) and another test (outside the pressure bulkhead) could be detected from the surface.
http://www.fire.tc.faa.gov/pdf/na68-7.pdf They were able to detect it to about 3000 yards at the surface (where the noise is worse).
And here's a fascinating patent, that mentions that the ULB's for KAL Flight 007 that crashed in '83 in the sea of Japan were never found and it mentions..."Recovery of the KAL 007 flight recorders may have been severely hampered by intentional acoustic jamming." It's a patent by Sandia Research for ULB's that are about the same size as the current version but is lower powered, with longer range, undetectable (signal is weak and below the noise floor) unless the correct code is known (spread spectrum), and can be uniquely identifyable to a specific aircraft/torpedo/etc.. Apparently it was developed for the military to use on experimental underwater toys in case they got lost and they wanted to be able to find them first before another nationality could find it. The ULB's would require only minor changes but the ping receivers would need some extra signal processing abilities.
Patent US4951263 - Spread spectrum underwater location beacon system - Google Patents