ACCURACY IN MODEL SEISMOLOGY

The most common technique currently used in model seismology makes use of a fixed source of elastic waves and a single receiver. This receiver is moved from position to position on a model in order to determine the relative spatial response to the source excitation. Inaccuracies occur due to the difficulty in obtaining the same receiver‐model coupling at each position. This coupling problem can be solved through the use of two source positions, the simultaneous use of two receivers, and the application of a general symmetry condition to the source‐receiver, and model geometries. Basic relations are established for this method. Independently of whether displacement, velocity, or acceleration detectors are used as receivers, these relations yield relative amplitude and phase spectra of the displacement at two different points on the model. These basic relations, and thus the validity of the method, are established by experiment. When a receiver is coupled to a model, the incident pulse is scattered as it passes the receiver position. This effect is negligible when bimorph receivers are used with two‐dimensional, steel‐sheet models. Edge receivers, used in conjunction with the same type models, produce significant, frequency‐dependent scatter. Quantitative results for this latter type of receiver‐model combination are obtained by experiment. These results show the amplitude‐spectrum of the displacement pulse to be attenuated by as much as 41 percent. Over the entire frequency range treated, the attenuation is never less than 4 percent. If it is necessary to use scattering receiver‐model combinations, the technique described for obtaining improved accuracy can be modified to correct for this effect.