Phased array antennas can produce a directional signal, known as "beamforming". The principle is they all act together, but out of phase (delayed) so that the "peaks" in the signal from each individual antenna, taking into account the propagation characteristics, are aligned in the desired direction or location. This can cause the signal from each antenna to act synchronously in that spot, but not in other spots, where they are out of phase and average out to noise. As for precisely how this works in the ionosphere, I don't know, but this is a common method in radio.

It is possible, perhaps likely that the effect is not very visible unless the observer is near the target, because most other places the energy is cancelling out. It's possible that the reason it showed up in the Japan data is because the sensor is in Japan.

If you have the ability to direct energy into the earth at a pinpoint location, it's not a stretch that you could trigger an earthquake, by loosening up a sticking point and allowing it to slip. Triggering an earthquake would require far less energy than is released, so the argument that you would notice an earthquake's worth of energy going to the location is not a relevant point. The energy is already there, pent up, waiting for the fault to overcome static friction.

So, I have no idea if that's what happened here, but it's not a priori impossible, unless there's a reason why these principles don't apply in this instance.

If this were the case, you would expect other "nodes" - places where the signal also adds constructively, for example, there might be another point in the direct opposite direction.