WAVES REFLECTED AT THE " SURFACE " OF THE EARTH : P ' P ' P ' P ' By

Phases of the 600 km. deep earthquake of April 16, 1957, which have not been reflected at the earth's surface show very sharp beginnings arriving within the limits of error of a few seconds at the cMculated times. However, phases which have been reflected at the surface, for example PP, pP', pPP, have more or less emergent beginnings which arrive between a few seconds and as much as 30 seconds too early. These emergent beginnings are followed by an impulse which arrives at nearly the calculated time. In agreement with earlier findings, it is concluded that the emergent beginnings of such waves are caused by waves which have been reflected at the Mohorovi6id discontinuity or other discontinuities below the surface. Observations of such early emergent arrivals also include waves of the types P 'P ' , PKSP' , P ' P ' P ' which have passed twice or more through the mantle and the core. P~P'P'P' has been observed for the first time. The travel times of P'P'P 'F and the coefficient of absorption calculated from P ' P ' P ' P ' and P' agree within the limits of error with the expected values; however, these limits of error are so large that the results cannot be used to improve earlier findings. The periods of P ' P ' P ' P ' which has traveled about 50,000 km. through mantle and core do not show a marked change from those observed in P waves which have traveled only a few thousand km. through the mantle. MATERIALS AND METHODS IT HAS been realized relatively early that in seismograms waves which have been reflected at the earth's surface may be preceded by waves reflected from discontinuities in and below the earth's crust. Gutenberg (1914, p. 183; p. 18 of reprint) found that PP frequently arrives about 10 seconds earlier than it was calculated. He pointed out that such waves with too early beginnings may be reflected at a discontinuity below the surface, and he calculated about 40 km. as its depth. Jeffreys (1926, pp. 344, 348) concluded that for reflection in continents "the wave reflected at the base of the granitic layer is probably stronger than that reflected at the outer surface when the observing station is less than 35 ° from the epicenter, and weaker, if the distance is greater." Such calculations assume constant velocity in each of two layers, extending to infinity from the boundary. From more recent calculations (Gutenberg, 1944, fig. lb, p. 95) it appears that under conditions discussed in the present paper the amplitudes of longitudinal waves reflected at the Mohorovi~i6 or similar discontinuities cannot be expected to be much greater than about 10 per cent of the amplitudes of the incident waves if the waves are short relative to the thickness of the crust. For longer waves there is no useful theory. Travel times of waves reflected at the Mohorovi~id and other discontinuities should be shorter than the times of those reflected at the surface by amounts depending on the usually varying depth of these discontinuities and the angle of incidence. Dipping or irregular form of these discontinuities must result in further complications for the travel times and amplitudes of such "early" reflected waves. Moreover, Gutenberg (1914, p. 183) had already pointed out that the latitude of the point of reflection of PP may produce differences amounting to as much as 6 Manuscript received for publication July 13, 1959. [71] 72 B U L L E T I N OF T H E SEISMOLOGICAL SOCIETY OF AMERICA seconds in the observed travel times. We can therefore expect ra ther wide scattering in the travel times of the early arrival of such waves reflected a t the ear th 's "surface," e.g., for PP, pP, pP ' , p tp , , SS, etc. These differences are even greater if there are multiple reflections, e.g., for pPP, sPP, P ' P ' P ' , SPS, etc. For an investigation of waves reflected a t the ear th 's surface, records of the earthquake of 4 b 04 m 04 ~, April 16, 1957, have been used. I t s epicenter has been found to t . f . -20 minute~ / h=6OOkm ~ ' -19 /X 6 X == /~o "x