During the last four years of U.S. research into theoretical seismology, effective techniques of computing synthetic seismograms have come into quite general use. Good agreement between a synthetic record and an instrumental record of ground motion (the raw data of seismology) implies a good understanding of the seismic source; of Earth structure; and of the theory of wave propagation relevant to the frequency content of the data. At periods T greater than about twenty seconds, there is now in general terms a fairly complete understanding of all the details in a seismogram, provided one ignores the (usually) minor effects of lateral variation of Earth structure. For 2展开▼
