BROAD-BAND MODELING OF REGIONAL SEISMOGRAMS - THE BASIN AND RANGE CRUSTAL STRUCTURE

Three-component broad-band displacement seismograms with paths samplin g the Basin and Range province are studied to constrain the crustal st ructure. To find an average model that fits the data in both absolute time and waveform, we generate broad-band reflectivity synthetics and conduct sensitivity tests on different parts of a layered crustal mode l, where only a few layers are involved. Generalized rays are used to help identify the various phases. It proves useful to decompose a regi onal seismogram into segments so that the impact of model parameters o n each segment can be clearly identified. Thus, for mid-crustal earthq uakes, it is established that the top crustal layer controls the Rayle igh wave, the Airy phase, in shape over the range from 300 to 600 km, and the crustal layer just above the source depth controls its timing. The P-nl waves, the P-n and P-L portion, are controlled in broad-band character by the mid-crust while the top layer contributes to their l ong-period motion. These crustal parameters control the tangential mot ion similarly. The SV wave, the segment between the P-nl wave and the Rayleigh wave, is mostly controlled by the shear velocity of the lower crust. In judging the goodness of fit between the array observations and synthetic waveforms, we allow individual data segments to shift re lative to the 1-D synthetics by a few seconds to account for some late ral variation. The amount of time shift is found by the cross-correlat ion in displacement between the data segment and the synthetics. Apply ing these tests in a forward modelling approach, we find that a simple two-layer crustal model is effective in explaining this data set. In this model, the main crustal layer has P and S velocities of 6.1 km s( -1) and 3.6 km s(-1), similar to those found by Langston & Helmberger (1974). A surface layer of thickness 2.5 to 3.5 km is required to fit the Rayleigh waves. The refined model can be used as a reference model for further studies in this region.