From 3-Hz P Waves to 0 S 2: No Evidence of A Slow Component to the Source of the 2011 Tohoku Earthquake

In the hours following the 2011 Honshu event, and as part of tsunami warning procedures at the Laboratoire de G,ophysique in Papeete, Tahiti, the seismic source of the event was analyzed using a number of real-time procedures. The ultra-long period mantle magnitude algorithm suggests a static moment of 4.1 x 10(29) dyn cm, not significantly different from the National Earthquake Information Center (NEIC) value obtained by W-phase inversion. The slowness parameter, is slightly deficient, but characteristic of other large subduction events such as Nias (2005) or Peru (2001); it remains significantly larger than for slow earthquakes such as Sumatra (2004) or Mentawai (2010). Similarly, the duration of high-frequency (2-4 Hz) P waves in relation to seismic moment or estimated energy, fails to document any slowness in the seismic source. These results were confirmed in the ensuing weeks by the analysis of the lowest-frequency spheroidal modes of the Earth. A dataset of 117 fits for eight modes (including the gravest one, (0) S (2), and the breathing mode, (0) S (0)) yields a remarkably flat spectrum, with an average moment of 3.5 x 10(29) dyn cm (*/1.07). This behavior of the Tohoku earthquake explains the generally successful real-time modeling of its teleseismic tsunami, based on available seismic source scaling laws. On the other hand, it confirms the dichotomy, among mega-quakes (M (0) > 10(29) dyn cm) between regular events (Nias, 2005; Chile, 2010; Sendai, 2011) and slow ones (Chile, 1960; Alaska, 1964; Sumatra, 2004; and probably Rat Island, 1965), whose origin remains unexplained.