I'm pretty sure this is what's happening. From past experience with high voltage arcs (admittedly at 60hz, not ~1 MHz), I don't think we're seeing more than a few hundred watts of power in that arc. Also, if this tower was actually the one transmitting that signal (presumably 50kW is the transmitter power) they'd shut it off or switch to a different antenna for maintenance on this one. If you shorted the output of the transmitter at the point where it's normally driving a resonant load, the best possible case is that it would trigger some kind of automatic shutdown on the final amplifier.