After the last post, I was thinking about that dominant 20-30 minute periodicity in the tidal gage responses to the tsunami waves on the west coast of the U.S., reflecting the time difference between tsunami peaks. That is one of the classic hazards with tsunamis: the wave trough may come in first, causing a beach to be emptied of water, and then when people walk out further away from shore to explore, the peak arrives half a period of time later, and people are trapped.
My visual memory of the tide gage data from the Japanese coast showed just an initial impulse. But when I looked at it again (see below, a blow-up from the last blog), one can see the initial impulse at about 30 minutes after the quake, and then a series of peaks for another 1-2 hours, at intervals of about 3-4 minutes.
Higher frequencies are attenuated more quickly with distance, as also happens for the spectral content in seismic surface waves, so by the time the waves reach the U.S. west coast, lower frequencies dominate.