auv-ee, gee, thanks for the numbers. They strongly suggest the 1700 meters range quote would be extraordinarily conservative. We start with 160.5 dBpa* at 1 meter and end at 30 dBps (18dBpa + some serious margin). That's 130dB of attenuation needed. Apparently we can expect 10dB from the properties of water. So that's a million to 1 range distance ratio before it's not detectable. Somehow I do not think the surface falls outside a thousand km from the emitter since the depth is single digit km.
3900 meters to 300 meters is 3600 meters. Call it 4000 meters and allow some modestly generous slant range. (That's nearly 2000 meters wide swath at the surface. 888*2 actually.) That would be 72dB loss or a level of about 88dBps looking at a 30dBpa threshold. Somehow a nearly 60 dB above threshold signal should blow the sonar operator out of his seat if he could hear it.
So unless somebody introduces new loss sources I withdraw my comments about the submarine not having a chance of hearing the pingers. And I start wondering what in heck happened to them that they BOTH failed to work.