The impedance of free space is more like 377 ohms, not 120; the impedance of twisted pair as used in ethernet cable is close to 120 ohms.
Maximum power transfer does, indeed, occur when the source and load impedance match, but this is not necessarily a desirable goal as it means that half the source voltage is dropped across its own internal resistance/impedance.
The tendency for corona/arcing to occur at a given voltage depends mainly on the gap between the electrodes, of course, as well as the pressure of the gas (if any) and its tendency to emit secondary electrons. This is codified in an equation called "Paschen's Law" and it is surprisingly accurate for a wide range of gases, pressures and electrode distances (IIRC, it does break down - pardon the pun - at really tiny gaps). The graph produced is roughly in the shape of an inverted parabola with a minimum arcing voltage in the range of 1-10 Torr for most gases (where 760 Torr = 1 Atmosphere = 1 Bar (more or less)).
Corona is, essentially, a low current (read: high impedance) "glow" discharge and as such won't be affected too much by the source impedance. Arcing, however, tends to be a much lower impedance type of discharge and so if the source impedance is too high then I would expect that arcing would stop almost immediately after it starts.