Two-dimensional linear elastic finite-difference calculations were performed for a two-dimensional geologic model of Yucca Flats, Nevada Test Site, Nevada. The calculations were used to produce synthetic teleseismic P-wave seismograms for explosive line sources in Yucca Flats. P-wave coda (in the first 5 sec) is observed to be highly dependent on take-off angle for the teleseismic synthetics. P-wave coda also varies with the position of the source in the valley structure and may produce variations in the individual station teleseismic P-wave mb magnitude of up to 0.3 magnitude units. However, these magnitude variations should be substantially reduced by averaging over stations at multiple azimuths.
The reverberant coda appears to rise from scattered modal waves that are initially excited in the low-velocity near-surface structures of the Yucca Flats deposits of alluvium and tuff. Scattering of the waves occurs at offsets in the basement structure and at the sides of the valley.
The combined effects of scattering, source function, intrinsic attenuation, and instrument response serve to obscure the P+pP spectral scalloping that is expected from a linear model. This loss of spectral resolution is the product of P-coda filling in the P+pP interference notches and the lengthening of the initial P-wave source time function by a convolution of the source time function, intrinsic attenuation operator, and instrument response. Therefore, short time windows (1 to 2 sec) that do not include P-coda energy do not have sufficient resolution to reliably detect the P+pP interference notches.