In this other thread, I was ready to believe the seemingly plausible assertion that "For ultrasonic transmission to work in water, you must use a transducer designed for working in water. They are quite expensive," and I hoped that the OP of that thread would do the tests I recommended, to prove the point, one way or the other.
He or she didn't do those tests, but it turns out that at least four others have, so there is good evidence that some ultrasonic transducers intended for air do, in fact, work underwater.
And, judging from some of the hobbyist's enthusiastic comments, they work well enough for their purposes.
Also, the CoconutPi kids measured depths up to 7 meters, not 160 cm. That, and the 50 feet claimed by one of the other experimenters, seems in the realm of "useful," but of course what is "useful" depends on one's goals.
And, if a $15 JSN-SR04T will do the job for some hobbyist, great. (As you noted, the only example using that particular sensor underwater was not entirely persuasive, and it would be interesting if someone, like the OP of this thread, could corroborate the findings...his/her current "0" could be because of too shallow depth.)
The one true sonar sensor you've identified lately (that I've noticed) is the "Airmar DST800 underwater transducer (cost about $270)" used in the cool Sea Renderer project that you mentioned. Which other sensor, costing less than the $99 MaxBotix and easy to use with the Arduino, did you have in mind?