That sounds pretty similar. The stuff we had in the lab could be plugged in directly to the wall and it would get very hot. They hooked it up to a resistor so they could control the heat. I supposed that would also work for consumer water pipe heaters.
If you go with something that produces radiant heat instead, one way to avoid the problem of hot air is to have the heating element above the bow. The hot air will rise up and away, while the actual infrared light rays which provide the radiant heat would still travel straight to the bow.
This was going to be my suggestion too. What better than a lamp*, if you want radiation of heat (and some light)? *Halogen or something hot, not an LED.