Echo return will be a no-go I think. The pipe is very thin and has perforations, which will increase the sound absorbing ability of the pipe walls. At 100 feet depth it will be like trying to get a sonic echo from a 3" object in an open field 100+ feet away (using a 3" speaker as your sound source). Actually it's worse than that, because of the other things in the pipe.
The resonant idea may pay off. You could sweep it through a freq range to suit the expected resonant freq of the pipe height with a microcontroller and look for some sort of "peaking" effect.
If the pump empties the bore, and then it fills back up, you could use the pipe and pressure sensor idea someone suggested like they use in a washing machine. That system works fine provided every now and then the water levels drops low enough to clear the bottom of the pipe so it re-equalises itself to atmospheric pressure, then will give an accurate reading when the bore fills back up. That could be a real good solution.
The resonant idea may pay off. You could sweep it through a freq range to suit the expected resonant freq of the pipe height with a microcontroller and look for some sort of "peaking" effect.
If the pump empties the bore, and then it fills back up, you could use the pipe and pressure sensor idea someone suggested like they use in a washing machine. That system works fine provided every now and then the water levels drops low enough to clear the bottom of the pipe so it re-equalises itself to atmospheric pressure, then will give an accurate reading when the bore fills back up. That could be a real good solution.