I guess that possibility cannot be ruled out all together Gary. However, it would seem to be a less likely explanation than ammonia volatilization. As James said above somewhere, you would expect denitrification to operate more effectively in a submerged filter where there are more anoxic areas. Anoxic areas can exist anywhere - even in floating detritus particles only a millimeter or so in size. But for denitrification to do any good, the process has to proceed all the way to the end point of nitrogen gas. It is a sequential process where nitrate (NO3) is converted to nitrite (NO2), then nitric acid (NO), then nitrous oxide (N2O), then finally nitrogen gas (N2). Nitrogen gas is very inert (the atmosphere is full of it) and readily volatilized. However, if the anoxic zone is disturbed before the nitrogen gas end point is reached, the process reverses itself, aerobic nitrification takes over again, and the partially reduced nitrogen is converted back to nitrate. In a TT you would expect the anoxic zones to be small in size and transitory making it difficult to get all the way to the denitrification end point.
There is another type of denitrification where nitrate is converted directly to ammonia. This process takes a special microbe which is less common than the typical anaerobe looking for a source of oxygen to scavenge. I guess it is possible that something about the TT could favor this specialized anaerobe. But, then you are back to ammonia volatilization as being the primary avenue for off-gassing nitrogen.
-steveh